xvEPA
United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Research Triangle Park NC 27711
EPA-600/7-78-213
November 1978
Reduction of
Nitric Oxide with
Metal Sulfides
Interagency
Energy/Environment
R&D Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide-range of energy-related environ-
mental issues.
EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-78-213
November 1978
Reduction of Nitric Oxide
with Metal Sulfides
by
P.P. McCandless and Kent Hodgson
Montana State University
Department of Chemical Engineering
Bozeman. Montana 59717
Grant No. R800682
Program Element No. 1NE624
EPA Project Officer: J. David Mobley
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
Abstract
This research program was initiated to determine the technical feasi-
bility of using metal sulfides for the chemical reduction of NO to N_.
Nineteen different metal sulfides were investigated for this use in an
initial screening using a test gas of pure NO. Although most sulfides inves-
tigated resulted in some NO reduction, BaS, CaS, SrS, and FeS were judged to
be the most promising. Several catalysts were found which reduced the tem-
perature at which the reduction reaction proceeds by as much as 200°C. A
further temperature reduction was obtained by impregnating the sulfide and
catalyst on high surface area supports of activated alumina or molecular
sieves. The most promising catalysts were NaF, NiCl , and FeCl .
Continuous electrobalance studies showed that O reacted with the sul-
fides at rates higher than with NO at higher temperatures but at 300°C CaS
mixed with NaF reacted with NO and not O . However, a test of this mixture
in a tubular reactor at 300°C resulted in no reduction of NO from a synthetic
flue gas stream.
All combinations of the most promising sulfides and catalyst were tested
for NO reduction in a tubular reactor using a synthetic flue gas containing
1000 ppm NO and 1 percent O . The capacities of the six best were FeS-FeCl
> SrS-NaF > CaS-NaF > BaS-FeCl > FeS-NiCl > CaS-FeCl and ranged from
0.0372 to 0.0134 grams NO reduced per gram of initial sulfide present. Capa-
cities of 0.91 and 0.76 were obtained when using 5% CaS (only) impregnated on
alumina and molecular sieves, respectively. It was concluded that these sul-
fides can reduce NO in the presence of O but more research is required to
establish the economic feasibility.
This report was submitted in fulfillment of Grant No. R800682 by Montana
State University under the (partial) sponsorship of the Environmental Protec-
tion Agency. Work was completed as of May, 1978.
ii
-------
CONTENTS
Abstract
Figures
Tables
Abbreviations and Symbols
Acknowledgment
Executive Summary ' lx
I. Introduction 1
Objectives and Scope of Work 1
Background 1
Thermodynamic Study 4
II. Conclusions 6
Thermodynamic Study 6
Preliminary Promoter Studies 7
Preliminary Studies of Dispersed Sulfides on High
Surface Supports 7
Kinetic Studies 7
Preliminary Tests Using a Synthetic Flue Gas 8
III. Recommendations 9
IV. Experimental Approach 11
Apparatus and Procedure 11
Analyses 17
V. Experimental Results 20
Preliminary Screening 20
Discussion of Preliminary Results 25
Preliminary Tests of CaS Impregnated on Supports 25
Electrobalance Studies 31
Rate of Reaction of Metal Sulfides with NO and 0 48
References 82
ill
-------
FIGURES
Number
1. Flow Diagram of Reactor System 12
2. Details of Flow Reactor 13
3. Diagram of Electrobalance Reactor System 14
4. Electrobalance Reactor Details 16
5. Schematic of the Apparatus used with the Tubular
Reactor for Synthetic Flue Gas Tests 18
6. Break Through Curves for CaS Pellets Promoted with NiCl ... 32
7. Reaction Data for a Harshaw Pellet at 390°C 34
8. Reaction Data for a Harshaw Pellet at 410°C 35
9. Reaction Data for a Harshaw Pellet at 437°C 36
10. Reaction Data for a Harshaw Pellet at 438°C 37
11. Reaction Data for a Harshaw Pellet at 468°C 39
12. Reaction Data for a Harshaw Pellet at 493°C 40
13. Reaction Data for Five Harshaw Pellets at 442°C 41
14. Reaction Data for a Harshaw Pellet using Oxygen 42
15. Reaction Data for a Harshaw Pellet at Low Flow Rates 43
16. Reaction Data for a Linde Sieve at 392 °C 44
17. Reaction Data for a Linde Sieve at 410°C 45
18. Reaction Data for a Linde Sieve at 438°C 46
19. The Percent NO and 0 Removed by 2 grams of
CaS/NaF at 400¥C 54
20. The Percent NO and O Removed by 2 grams of
CaS/NiCl at 4^0°C .7 55
21. The Percent NO and O Removed by 2 grams of
CaS/CoCl at 4§0°C 56
22. The Percent NO and O Removed by 2 grams of
CaS/FeCl2 at 4&)0C .7 57
23. The Percent NO and 0 Removed by 2 grams of
CaS/Fe O at 48o°C .7 58
24. The Percent NO and 0 Removed by 2 grams of
SrS/NaF at 400*C 59
25. The Percent NO and O Removed by 2 grams of
SrS/NiCl2 at 4&)0C 60
26. The Percent NO and O Removed by 2 grams of
SrS/CoCl2 at 48o°C .7 61
27. The Percent NO and O Removed by 2 grams of
SrS/FeCl at 4$0°C 62
28. The Percent NO and 0 Removed by 2 grams of
SrS/Fe 0 at 4&)°C .7 63
29. The Percent NO and O Removed by 2 grams of
BaS/NaF at 400*C 64
30. The Percent NO and O Removed by 2 grams of
BaS/NiCl2 at 4$0°C .7 65
31. The Percent NO and O Removed by 2 grams of
BaS/CoCl2 at 4&)0C .7 66
iv
-------
Number Page
32. The Percent NO and O Removed by 2 grams of
BaS/FeCl2 at 4&00C .7 67
33. The Percent NO and O Removed by 2 grains of
BaS/Fe 03 at 4$00C .7 68
34. The Percent NO and O Removed by 2 grams of
FeS/NaF at 400*C . . 7 69
35. The Percent NO and O Removed by 2 grams of
FeS/NiCl at 4&)0C .7 70
36. The Percent NO and O Removed by 2 grams of
FeS/CoCl at 4$00C .7 71
37. The Percent NO and 0 Removed by 2 grams of
FeS/FeCl2 at 4^0°C . 7 72
38. The Percent NO and O Removed by 2 grams of
FeS/Fe O at 4&>°C . 7 . 73
39. The Effect of SO in the Feed Gas Stream on
NO Removal at 400°C with CaS/NaF 74
40. The Effect of HO on NO Kemovai by CaS/NaF
at 400°C ... 7 ... * 75
41. The Effect of HO on the Removal of NO by 2 grams
of SrS.NaF at 400°C * 76
42. The Effect of O Concentration on NO Removal 77
43. The Removal of NO by Harsnaw Pellets and Linde
Molecular Sieves Impregnated with CaS 79
-------
TABLES
Number Page
1. Calculated Free Energy Change and Heat of Reaction for
the Reduction of NO with Various Metal Sulfides 5
2. Preliminary Tests of Various Metal Sulfides for NO
Reduction 21
3. Catalyst Studies for NO Reduction Using Various
Sulfides Test Gas Pure NO 22
4. Preliminary Tests for SO Formation 24
5. Reduction Using 2.5% NO in He 24
6. Physical Properties of Support Material 26
7. Reaction Conditions 33
8. Average Rates of Reaction vs. Temperature 47
9. A Summary of the Rate of Weight Gain or Loss for
the Various Sulfides Reacting with NO and 0 49
10. Reaction Rates of NO and O with Various Metal Sulfides .... 50
11. The Relative Rates at Which NO and O React with the
Various Sulfides bO
12. Rate of Reaction of CaS at Various Temperatures Using
Different Promoters 51
13. The Weight Change with Tune of Various Promoters
Reacting with NO or O bl
14. The Rate at Whicn NO Reacts with the Various Metal
Sulfide Chemical Promoter Mixtures 52
15. The Capacities of the Metal Sulfide-Catalyst Mixtures
and of the Metal Sulfides 53
vi
-------
LIST OF ABBREVIATIONS AND SYMBOLS
AgF
Ag S
Ag,so.
"k
A12S
A1^
BaS
BiS
CaCl
CaO
CaS
CaSO.
CaSifl
CdS
CdSO.
CO 4
CoCX
CuCl
CuCl
CuS
Cu S
CuSO
Fed*
Pe2°3
FeS J
K CrFi
K^FeFi
K^ZrFi
argentic fluoride
argentic sulfide
argentic sulfate
alumina
aluminum sulfide
aluminum sulfate
barium sulfide
bismuth monosulfide
calcium chloride
calcium nitrate
calcium oxide
calcium sulfide
calcium sulfite
calcium sulfate
calcium metasilicate
cadmium sulfide
cadmium sulfate
carbon monoxide
carbon dioxide
cobaltous chloride
cuprous chloride
cupric chloride
cupric sulfide
cuprous sulfide
cuprous sulfate
ferrous chloride
ferric oxide
ferrous sulfide
ferrous sulfate
hydrogen
water
hydrogen sulfide
mercurous sulfide
potassium chromium fluoride
potassium iron fluoride
potassium zirconium fluoride
V
MeS
MeSO
MnS
MnSO
,
NO
NO
NaF
Nal
Na^O
NH
NH^OH
NiCl
NiS
PbS
PbSO
PtC1
SO
Sr^
SrSO
ThS
T£ 5
WS2
Zns
ZnSO
ferrous sulfate
metal sulfide
metal sulfate
manganous sulfide
manganous sulfate
molybdenum disulfide
nitrogen
nitric oxide
nitrogen oxide
nitrous oxide
nitrogen trioxide
nitrogen tetroxide
nitrogen pentoxide
sodium chloride
sodium fluoride
sodium iodide
sodium oxide
ammonia
ammonium hydroxide
nickelous chloride
nickel monosulfide
oxygen
palladium chloride
lead sulfide
lead sulfate
platinum dichloride
sulfur
antimony trisulfide
silica
sulfur dioxide
strontium sulfide
strongium sulfate
thorium sulfide
thallium sulfide
tungsten disulfide
zinc sulfide
zinc sulfate
cm
g
hr
mg
min
cubic centimeters
grams
hours
milligrams
minutes
ml milliliters
ppm parts per million
SCR selective catalytic
reduction
sec seconds
vii
-------
ACKNOWLEDGMENT
Many individuals were very helpful in carrying out this project.
Montana State University Chemical Engineering students. Jerry Bowman, Russell
Erickson, John Mclntyre, and Richard White, all contributed greatly to the
experimental investigation. In addition, the invaluable assistance of Silas
Huso and Jim Tillery, who constructed much of the laboratory equipment, is
gratefully acknowledged.
\rili
-------
EXECUTIVE SUMMARY
This report presents the results of a preliminary investigation of the
novel use of metal sulfides for the chemical reduction of NO to N with the
objective of determining the technical feasibility of using metal sulfides
to control NO pollutants from stationary sources.
X
In an initial screening, nineteen different metal sulfides were investi-
gated for this use. In tests of the bulk powdered sulfides, using pure NO,
all but one reduced NO to N at temperatures varying between 90 and 650°C.
However, there was a weight loss using many of the sulfides indicating an un-
desirable side reaction was occurring, probably the formation of SO . Most
of the tests also showed that at least some metal sulfate was formed. BaS,
CdS, CaS, Cu S, PbS, SrS, ZnS and NiS all gained weight but the alkaline
earth sulfides were judged to be the most promising because of the stability
of the sulfides and sulfates and because previous research has shown that the
sulfate can easily be reduced to the sulfide. Due to its availability, most
of the subsequent tests used CaS.
The temperature at which the reduction reaction proceeds was signifi-
cantly reduced (by as much as 200°C) by intimately mixing the powdered sul-
fide with suitable catalysts. Of the materials tested, NaF and the transi-
tion metal chlorides appeared to be the most active. The temperature was
also greatly reduced by impregnating the sulfide on a high surface area solid
support. Activated alumina appeared to be the best support since side reac-
tions producing SO were observed with silica, silica alumina, and molecular
sieve supports. Tne temperature for reaction was further reduced by impreg-
nating both the sulfide and catalyst on the pellets. Kinetic studies using
a recording electrobalance showed that the rate of reaction of the sulfides
with O was generally greater than that with NO but that certain catalysts
selectively increase the rate of reaction with NO and indicate that some
sulfide-catalyst combinations will react with NO and not O at temperatures
below 300°C.
Finally, initial tests using a synthetic flue gas containing 1,000 ppm
NO, 1% O , and 18% CO (balance N ), showed that NO was rapidly reduced by
CaS at temperatures aBove 600°C. The reaction temperature was reduced to
about 400°C by mixing the metal sulfides with a catalyst.
All combinations of BaS, CaS, SrS and FeS with NaF, NiCl2, CoCl2, FeCl2
and Fe 0 were tested at 400°C using the synthetic flue gas and the capacity
for NO2 reduction calculated. Capacity is defined as the weight of NO re-
duced per unit weight of sulfide initially present from the start of the run
until the exit concentration exceeds 600 ppm. Of the combinations tested,
the six best were: FeS-FeCl > SrS-NaF > CaS-NaF > BaS-FeCl, > FeS-NiCl >
CaS-FeCl . For these combinations capacities varied from 0.0372 to 0.0134
grams NO reduced per gram of metal sulfide intially present when tested in a
packed bed tubular reactor. Using the same test conditions 5 percent CaS
(only) impregnated on a Harshaw activated alumina and a Linde Molecular sieve
ix
-------
had capacities of 0.76 and 0.91 grams of NO reduced per gram of metal sulfide
initially present. This indicates that capacity can be greatly increased by
making the sulfide more available for reaction. For comparison, assuming the
reaction:
CaS + 4NO •* CaSO + 2N
the maximum capacity for CaS would be 1.66 grams NO/gram CaS. Capacity was
greatly reduced by increasing the O concentration to 10 percent.
At 400°C the presence of 2 percent water vapor reduced the rate of reac-
tion, probably by interfering with the action of the catalyst. At higher
temperatures, where the reaction probably is not catalytic, there was no
effect. SO and CO had no effect on the reduction reaction.
^ £•
It is concluded that calcium sulfide is capable of reducing NO even in
the presence of O . However, more research is required to better define an
optimum system ana to establish the economic feasibility of using this
process to control NO emissions.
x
-------
SECTION I
INTRODUCTION
OBJECTIVES AND SCOPE OF WORK
The objective of this research program was to investigate the use of
metal sulfides for the reduction of nitrogen oxides with the aim of control-
ling emissions of these pollutants from stationary sources. The study in-
cluded an extensive laboratory investigation of various metal sulfides both
in bulk powdered form and sulfides impregnated on high surface area supports
for the reduction of NO; an investigation of the use of various materials to
catalyze the reduction reaction; and continuous recording electrobalance
studies so that rates of reaction could be measured.
BACKGROUND
Oxides of nitrogen are one of the most prevalent air pollutants with
present (1975) total annual emissions estimated at about 20 million tons per
year in the U.S. alone. They are major reactants in the formation of photo-
chemical smog and their emissions must be controlled because of their adverse
effects on human health and plant life (1).
Emission levels of nitrogen oxides will probably rapidly increase in the
future, even with the application of the best control technology currently
available because of the very rapid growth rate of combustion sources as
energy demands continue to rise. This is in contrast to emissions of SO and
particulate matter which are also produced in the combustion of fossil fuels.
Unlike NO. both SO and particulate emissions can probably be maintained at
present levels, or reduced, using present technology (2).
Of the current U.S. NO emission, about 55 percent originate from sta-
tionary sources and 93 percent of all stationary NO emissions are from the
combustion of fossil fuels for steam and/or electricity generation, or
space heating (3).
Emission levels from stationary sources are most frequently in the 50
to 1000 ppra range (8), but at the very high temperatures that will be pres-
ent in the combustor of an MHD generator the equilibrium NO concentration
may be as high as 3.25 percent (32,500 ppm) (4). Organically bound nitrogen
in the fuel may set a lower limit on the formation of nitric oxide, partic-
ularly if a high nitrogen containing fuel such as coal is used (5,6).
Nitric oxide (NO) is generated in all high temperature processes in-
cluding air by the direct combination of nitrogen and oxygen:
N2 + °2 J 2N°
-------
The kinetics of the N , O , NO reaction system is such that an equilibrium
composition is rapidly established at temperatures above 2000°K. NO becomes
unreactive at lower temperatures even though its decomposition is thermody-
namically favored by low temperatures. The decomposition reaction is very
slow at temperatures below 2000°K and after temperatures of about 1800°K and
below are reached very rapidly the NO concentration is essentially "frozen"
at the high temperature equilibrium value. Surfaces may exhibit a catalytic
effect for the decomposition reaction. Thus, the concentration of NO
appearing in an effluent gas from a combustion process depends on the combus-
tion process, the fuel-oxidizer mixture, and the time-temperature-surface
contact history of the gas as it goes through a particular combustor system.
Nitric oxide (NO) is the species formed in the high temperature combus-
tion but it is quite unreactive at lower temperatures. At ambient tempera-
tures, it is oxidized slowly to the more reactive and irritant nitrogen
dioxide (NO ) by oxidizing agents (ozone, O ) in the atmosphere. NO is the
main cause of the highly visible red-brown color of smog. The relative
amounts of NO and NO in a gas containing oxygen vary widely depending on
temperature and O concentration and the combination (NO + NO ) is usually
referred to as combined NO .
x
Nitric oxide (NO) and nitrogen dioxide (NO ) are the major pollutants
but the unstable NO , NO, NO, and NO forms also exist. Nitrous oxide
(NO) is a stable oxide which is formed oy the decomposition of nitrogen com-
pounds in the soil by bacterial action and the atmosphere contains about
0.5 ppm of this compound from natural sources. NO is also an intermediate
in thfe catalytic reduction of NO, and in some cases it may be a substantial
portion of nitrogen oxides emitted from chemical processes with nitric acid.
N.O is innocuous and considered non-polluting.
There are (at least) two general routes available for controlling NO
emissions from stationary combustion sources; combustion modification andx
flue gas treatment.
The release of NO in combustion gases can be minimized by several com-
bustion modification techniques, but not entirely eliminated. The most prom-
ising combustion modification methods include: load reduction, low excess
air firing, two stage combustion, off-stoichiometric firing, flue gas recir-
culation, water and steam injection, or a combination of these techniques
(7). However, many of these less costly control methods can easily produce
undesirable results such as reduced efficiency, heat transfer surface corro-
sion and increased CO and hydrocarbons emissions (2,8). Additionally, as
noted above, organic nitrogen compounds in the fuel may put a lower limit on
NO production from coal, and to a lesser degree with some other fuels (2,5,
6,i). For example, when an oil containing about 0.2% nitrogen is burned, the
amount of NO produced from the fuel is about equal to that of thermal NO (5)u
Although there is no thermodynamic hindrance to the decomposition of NO
to N and O at temperatures below 1000°K, apparently the energy of activa-
tion of the NO molecule is very high and this results in the decomposition
reaction being kinetically limited. Heterogeneous catalysts are available
that reduce this activation energy considerably, but reaction rates are still
too low to be practical for NO emission control, apparently because of an
unusually small pre-exponential term in the reaction rate constant (9).
-------
Since the decomposition of NO to N and O is too slow to be practical
at moderate temperature even with the best catalysts available, the hetero-
geneous catalytic reduction of NO to form N has been studied using a wide
variety of reducing agents (hydrocarbons, activated carbon, hydrogen, carbon
monoxide, ammonia) and catalysts with the aim of the chemical destruction of
NO in gas streams. In particular, reduction with ammonia shows promise since
it can reduce NO and NO without simultaneously reacting with oxygen between
the limits of 210 to 270°C. Above 270°C, ammonia will react with O to form
NO, below 210°C, it apparently forms ammonium nitrate (10). 2
Since 1973 selective catalytic reduction (SCR) with ammonia has been
used in Japan in many industrial plants for NO control (Ando et al., 1977).
These plants apparently use reaction temperatures of 300 to 450°C. Under
these conditions the presence of oxygen promotes the reactions and the re-
duction reactions can be represented by the following equations:
4NH + 4NO + 0 -»• 4N + 6H 0
J ft £ f,
4NH + 2NO + O •* 3N + 6H O
A large number of proprietary base-metal catalysts have been developed
to promote the reduction reaction with claims of over 90% NO reduction
being obtained using NH :NO ratios of about 0.8 to 1.4:1 wi£h less than
10 ppm NH in the treated flue gas.
Other technically feasible methods for NO control from stationary
sources include absorption or adsorption. However, since NO itself is quite
unreactive, absorption or adsorption usually requires a two stage process to
first oxidize the NO to either NO or NO followed by treatment of the oxi-
dized gas for NO removal. This greatly complicates the process because the
oxidation of NO to NO is a relatively slow reaction with the rate decreasing
with increasing temperature. To date, performance tests of the sorption
methods have ranged from unsatisfactory to partially successfull (10,11,12).
In considering the various methods for NO minimization and/or flue gas
denitrification that are available or in the development stage, certainly
the combustion modification techniques are the simplest and in most cases
more economical to use. However, usually this is only helpful in reducing
NO emission from new facilities. In most cases a modification of combustor
design is necessary to provide significant NO reduction. This is difficult
in existing plants but good results are obtained when the design modifica-
tions are incorporated into new plants. However, many of these less costly
modification techniques can produce such undesirable results as lower fuel
efficiency and increasing hydrocarbon and CO emissions. In addition,
stricter standards are likely to require flue gas denitrification to reduce
NO levels below what can be obtained by combustion modifications in some
cases (2). This is particularly true of coal fired plants because it is
difficult to reduce the NO in the flue gases from coal burning below
400 ppm by combustion modifications alone (5).
Of the flue gas denitrification techniques being used in Japan (where
strict emission standards require low emission levels), several modifica-
tions of selective catalytic reduction with ammonia are the processes most
widely used. The major problem with this process is plugging of the cata-
lyst with dirt, catalyst poisoning by dust components, SO and SO , and the
3
-------
formation of ammonium bisulfate which is very corrosive. In addition, since
ammonia injection ratios typically are NH :NO=1.1^1.4, large amounts of ammo-
nia are required. Indeed, in Japan it is estimated that about 400,000 tons
of ammonia per year would have to be consumed to attain ambient air standards
using the SCR method and this is equivalent to about 1/2 the total annual con-
sumption of nitrogen fertilizers in that country! This would appear to be a
socially unacceptable use of a valuable fixed nitrogen fertilizer when the
world food production is being limited by fertilizer shortages (12).
in any case, however, it is necessary and desirable to develop alter-
native flue gas denitrification processes as well as better combustion modi-
fication techniques as long as the consumption of fossil fuels continues to
increase (2,5,12).
The literature of 60 years ago (13,14) discusses the "Thiogen" processes
for the recovery of sulfur from SO according to the reactions
CaS + 2SO0 -»• CaSO, + S^
2 42
4CaS + 6SO -»• CaSO + 3S
Preliminary research by the authors indicated that the reaction
CaS + 4NO -*• CaSO. + 2N.
4 2
proceeds rapidly at temperatures above 450°C when pure NO is passed through a
reactor containing powdered calcium sulfide. Based on the promise of this
type of reaction for NO emissions control, a research program was initiated.
This report presents the results of a preliminary investigation of various
metal sulfides for reducing NO together with tests of various materials to
catalytically enhance the reaction.
THERMODYNAMIC STUDY
Thermodynamics gives a good indication of the potential of metal sul-
fides for NO reduction and so a brief study was made to compare the reduc-
tion reactions using various metal sulfides. Several possible side reactions
of CaS were also considered together with the thermodynamics of reducing NO
with hydrogen. Possible side reactions with other gaseous compounds which
may be present in flue gases were not considered. The study was limited to
systems for which data on both the sulfide and corresponding sulfate were
available. Approximate methods were used to calculate the effects of temper-
ature on the free energy of reaction where heat capacity data were not avail-
able. The results are summarized in Table 1.
Several interesting and important conclusions can be drawn from these
data. First, the reduction reaction of NO using metal sulfides are very
favorable at all temperatures of practicalxinterest and in particular at tem-
peratures below 1000°K. In addition, from a standpoint of free energy driv-
ing force, the use of metal sulfides compares favorably with reduction using
a conventional reducing agent such as hydrogen. This includes the reduction
of both NO and NO with the formation of either N and/or NO. In addition,
from a thermodynamic standpoint it appears that tne reduction of NO or NO
is favored over the oxidation of the sulfide with free oxygen and indicates
' that the formation of the sulfate is favored over the formation of SO. and
-------
CaO. However, the same data also show that these undesired side reactions
are also very possible (and indeed probable). Unfortunately, thermodynamics
only indicates which reactions are theoretically possible and gives no indi-
cation as to how fast the reactions will proceed or the relative rates of the
various reactions. The kinetics of the reactions must be determined from ex-
perimental data. Thus, experimental data are required to determine if prac-
tical use of these reactions can be made to control NO emissions.
x
TABLE 1. CALCULATED FREE
REDUCTION OF
ENERGY CHANGE AND HEAT OF REACTION FOR
NO WITH VARIOUS METAL SULFIDES
THE
A:
KCal/Mole
Reaction
CaS+4NO-»CaSO +2N
MnS-M NO->MnS04+2N
PbS+4NO+PbSO.+2N_
4 2
ZnS+4NO+ZnSO +2N
Cds+4NO-»CdSO +2N
CuS+4NO-»C uSO +2N
Ag S+2NO*Ag SO +2N
1/3 Al S +4NO->l/3 Al (SO ) .
4H +4NO-»4H 0+2N
CaS+1. 50 ->CaO+SO
CaS+2O ->CaSO
CaS+3NO->CaO+SO +1. 5N
CaS+2NO -»CaSO +N
CaS+8NO->CaSO.+4N_0
4 2
298°K
-284.
-261.
-253.
-243.
-242.
-229.
-220.
,+2N_ -289.
3 2
-301.
-102.
-201.
-164.
-226.
-268.
3
4
6
7
0
3
3
9
3
0
5
32
24
18
500°
-264.
-241.
-234.
-224.
-222.
-209.
-200.
-270.
-290.
- 98.
-184.
K
3
0
2
5
8
5
9
3
4
3
1
-158.88
-219.
-231.
60
80
Sulfide
1000°
-215.
K
7
-190.0
-186.
-177.
-174.
-160.
-152.
-221.
-263.
- 89.
-141.
-145.
-195.
-141.
1
0
8
2
4
1
2
1
1
39
91
60
1500°
-166.
-139.
-138.
-129.
-116.
-110.
-104.
-171.
-236.
- 79.
- 98.
-131.
-172.
- 51.
K
7
2
0
5
8
8
0
8
1
8
1
91
21
40
AHR
KCal/
Mole
Sulfide
298 °K
-313
-291
-282
-273
-270
-258
-249
-319
-317
-107
-227
-172
-243
-322
.5
.8
.3
.1
.8
.8
.3
.5
.5
.6
.5
.36
.30
.00
-------
SECTION II
CONCLUSIONS
The novel use of metal sul fides to chemically reduce nitrogen oxides to
elemental nitrogen was investiged in a comprehensive screening program. The
following important conclusions were made based on the preliminary experimen-
tal investigation.
THERMODYNAMIC STUDY
(a) A thermodynamic study showed that there are no thermodynamic limi-
tations on reactions of the type:
4NO + MeS -> MeSO^, + 2N0
4 2
2NO^ + MeS -»• MeSO, + N
2 42
where "Me" represents the various metal sulfides studied.
These reactions have very high negative free energy changes indicating they
are very favorable at all temperatures of practical interest. Conversion to
nitrogen would be virtually complete at equilibrium.
(b) The free energy changes for the sulfide reduction of NO compares
favorably with reduction using conventional reducing agents such as H or
carbon.
(c) However, the oxidation of the sulfides with oxygen is also thermo-
dynamically possible as well as other undesirable side reactions which result
in the formation of gaseous sulfur compounds.
(d) The reduction reactions using sulfides are highly exothermic.
PRELIMINARY SCREENING OF BULK POWDERED SULFIDES
(a) Nineteen sulfides were tested for NO reduction using pure NO as
the test gas. All but one resulted in the reduction of NO to N and/or NO
and in most cases at least some sulfate was formed. The temperature at
which N and/or NO was formed varied with the different sulfides between
90 and 550°C.
(b) A12S3' Sb2S3' BiSf CUS' FeS/ MnSf HgS' M°S2' K2S' WS2.' and TlS'
all lost weight during the reaction with NO indicating undesirable side
reactions were occurring.
(c) BaS, CdS, CaS, Cu S, PbS, SrS, ZnS, and NiS all gained weight dur-
ing reaction with pure NO. Of these, the alkaline earth sulfides, BaS,
-------
CaS and SrS, were judged to show the most promise, particularly CaS because
of the abundance of CaSO which can be reduced to the sulfide.
4
PRELIMINARY PROMOTER STUDIES
(a) Several materials were found to increase the rate of the reduction
reaction including K FeF CuCl, CaCl , NiCl , NaF, Nal, CoCl , and FeCl
when physically mixed with some of the sulfides. This resulted in the reac-
tion occurring at temperatures about 200°C lower than with the sulfide alone.
The investigation of possible catalysts was limited to materials on hand and
was by no means exhaustive.
(b) K3ZrF&, K3CrF6, FeSO4, Fe 0 , and Cr O were less active or did not
enhance the reduction reaction.
(c) Some of the catalysts appeared to result in the formation of SO at
higher temperatures.
PRELIMINARY STUDIES OF DISPERSED SULFIDES ON HIGH SURFACE SUPPORTS
(a) A very active form of CaS can be obtained by impregnating Ca(No )
on the high surface area solid such as activated alumina, silica or molecular
sieves; conversion of the nitrate to the oxide, and conversion of the oxide
to the sulfide with H S. This results in the NO reduction reaction occurring
at temperatures below 100°C.
(b) Impregnation of a catalyst on the support along with the sulfide
further lowers the reaction temperature so that reduction occurs at room tem-
perature.
(c) Oxygen also reacts with the high surface area sulfide pellets, even
at room temperature.
(d) The high silica and molecular sieve supports resulted in the for-
mation of SO at higher reaction temperatures.
KINETIC STUDIES
The following conclusions were made from kinetic studies of the rate of
reaction of the sulfides using a continuous recording electrobalance: (a)
The rate of reaction of 8.6% CaS impregnated on activated alumina varied from
1x10 to 1.8xlO~ mole NO removed/(hr)(gram pellet) at temperatures of 390
and 493°C respectively when tested with a gas mixture containing 2.5% NO in
helium. In most cases, the rate rapidly declined with time indicating signif-
icant diffusional resistance to reaction as the sulfide in the pellet is con-
sumed.
(b) Reaction rate measurement with powdered sulfides alone showed that
the rate of reaction of 0 with the sulfides is more rapid than with NO in
most cases. The ratio of the rates OVNO varied between 3 and 8 for CaS at
a temperature between 400 and 500°C.
(c) CaS mixed with NaF or NaCl as catalysts reacts only with NO at lower
temperatures. However, in a tubular reactor using a test gas containing both
-------
NO and O the rate was too low to detect a change in the NO concen-
tration.
(d) In tests of pure promoters, K FeF reacted with NO at all temper-
atures investigated and with O at 500°C. NaF and NaCl did not react with
either NO or O at temperatures of 400°C and below. NiCl and FeCl did
not react with NO or 0 at 300 or 400°C.
PRELIMINARY TESTS USING A SYNTHETIC FLUE GAS
(a) A 5 gram bed of unpromoted CaS reduced about 98% of the NO in a gas
stream containing 1000 PPM NO and 1% O at temperatures between 600 and
700°C at a gas flow rate of ifo cc/min.
(b) The capacities at 400°C of all combinations of BaS, CaS, SrS and
FeS mixed with NaF, NiCl , CoCl_, FeCl and Fe_O for NO reduction was deter-
mined using a synthetic flue gas. The gas contained 1000 ppm NO, 1 percent
O , 18 percent CO , with the balance N . The capacity is defined as the
weight of NO reduced per unit weight or metal sulfide initially present from
the start of the run until the exit combination exceeds 600 ppm. Of the com-
binations tested the six best were FeS-FeCl > SrS-NaF > CaS-NaF > BaS-FeCl
> FeS-NiCl > CaS-FeCl . For these combinations, capacities varied from
0.037 to 0.013 grains NO reduced per gram of initial sulfide present.
(c) Using the same test conditions as in (b), 5 percent CaS (only)
impregated on a Linde Molecular Sieve and on a Harshaw activated alumina had
capacities of 0. 91 and 0.76 grains of NO reduced per gram of CaS initially
present.
(d) The presence of 10 percent 0 in the gas stream greatly decreased
to capacity of CaS-NaF to reduce NO at 400°C.
(e) The presence of 2 mole percent of water vapor in the gas stream
greatly decreased the capacity for CaS-NaF and SrS-NaF at 400°C.
(f) SO and. CO had no effect on the capacity of CaS for NO reduction.
(g) NO can be reduced with CaS, BaS, SrS, or FeS in the presence of
1 percent O .
8
-------
SECTION III
RECOMMENDATIONS
The preliminary results with CaS mixed with certain catalysts indicates
that NO is rapidly reduced at temperatures above 400°C in the presence of O
althougn the sulfide also reacts rapidly with O under these conditions. How-
ever, the kinetic studies indicate that certain catalysts are selective for
NO; that is, the rate of reaction of NO with the sulfide is increased while
with O it is not. In these tests, CaS reacted with NO and not O at temper-
atures below 300°C when promoted with NaF. Further, only CaS has been tested
extensively although the kinetic studies indicated that BaS or SrS may be
better in the presence of 0 . As a result, it may be possible to find a
sulfide-catalyst combination which will rapidly react with NO but not O under
certain conditions and thus minimize the consumption of the sulfide witn O .
This may also be true for a metal sulfide-catalyst mixture impregnated on
high surface area pellets.
Also, the effects of gas composition, in particular O and HO, have not
been adequately investigated. In addition, the side reaction in which gase-
ous sulfur compounds are produced have not been investigated in any detail.
Only the formation of SO and H S have been observed under certain conditions.
The reaction mechanism for these reactions must be determined and the condi-
tions under which the reactions occur must be better defined. Thus, there
are many critical factors which must be investigated further before both the
technical and economic feasibility of using this process for pollution control
can be established. As a result, it is recommended that further work be car-
ried out including the following:
1) Further kinetic studies should be carried out to determine the rel-
ative rate of reaction of NO and 0 with promoted BaS, CaS, SrS, FeS, and
possibly ZnS and CdS. An extensive screening of possible catalysts should be
carried out in connection with this study. This would further indicate which
systems and conditions might result in the reduction of NO in the presence
of 0 with a minimum of oxidation of the sulfide with 0 . The use of a gra-
dientless reactor of the Berty or Carberry type (17,18) would be ideal for
this study.
2) Further tests in a bench scale flow reactor should be made using
the most promising sulfide-catalyst combinations and a synthetic test gas
with a composition more typical of power plant effluent gas streams. The
effects of temperature, sulfide-catalyst composition, space time, gas compo-
sition, etc., on NO reduction and the reaction of 0 should be determined.
3) Studies similar to that outlined in (2) should be made where the
sulfide-catalyst is impregnated on high surface area supports.
-------
4) A study whould be made of the regeneration of the spent sulfide-
catalyst.
The results of these studies would permit a preliminary economic evalu-
ation of NO pollution control using sulfides.
10
-------
SECTION IV
EXPERIMENTAL APPROACH
APPARATUS AND PROCEDURE
A simple semi-batch reactor system (continuous flow of the test gas to
the reactor but batch-wise sulfide reactant addition) was used for many of
the tests including initial screening of various metal sulfides, catalyst
tests, and more detailed preliminary tests of the most promising sulfides. A
schematic diagram of this reactor system is shown in Figure 1. The test gas
mixture (previously blended) from a high pressure cylinder was fed to the
reactor through a calibrated rotameter. For some tests the gas mixture was
first saturated with water vapor by sending the gas through a gas washing
bottle before it was sent to the reactor. The details of the reactor are
shown in Figure 2. Several size reactors were used at various times in the
research but all had the same basic design. The reactors were constructed
from 14 to 18 inch lengths of schedule 40 type 304 stainless steel pipe with
diameters of 1.5, 0.75 and 0.5 inch with appropriate inlet and outlet fit-
tings. The bottom half of the pipe was packed with stainless steel wire
rings to increase heat transfer and ensure that the gas entering the reaction
chamber was at the desired temperature. The top half constitutes the reac-
tion chamber and is contained between two porous stainless steel plates. A
thermowell made from 1/4" stainless steel tubing was mounted axially in the
pipe where thermocouples were mounted at three locations in the reaction
chamber. In operation the reactor was mounted in a 3 or 4 inch diameter
metal block with a slightly oversize hole through it. This block was wrapped
with three nichrome heating coils in ceramic beads and controlled by variable
transformers. This heater-block assembly was mounted in an outer container
which was filled with vermiculite insulating material. Both bronze and
stainless steel heating blocks were used but the bronze block proved to be
less satisfactory due to oxidation corrosion at higher temperatures. The
metal block insured near isothermal operation throughout the reactor length
and helped to eliminate both radial and longitudinal temperature gradients.
The gas outlet passed through a small water cooler and then through a tee
containing a silicone rubber sampling septum and thence to either a continu-
ous analyzer or a vent hood. Stainless steel fittings and stainless steel
tubing were used throughout the reactor system. A typical test was made as
follows: 2 to 20 grams of the powdered sulfide or sulfide-catalyst mixture,
or sulfide impregnated on high surface area support, was placed in the reac-
tor and the test gas fed to the reactor at a constant rate while the reactor
was heated. Samples of the exit gas were taken periodically in a gas tight
syringe and analyzed by gas chromatography.
Relative rates of reaction were determined with a Cahn R-100 continuous
recording electrobalance in the apparatus shown schematically in Figure 3.
The balance has a 100 gram capacity for sample weight plus container. Mechan-
ical tare capacity is 100 grams and the balance has three electrical weight
11
-------
EXIT GAS COOLER
*&
TI
Ml
ROTAMETEE
a ^
n
JST GAS
[XTURE
"**!
6
l\
\
f
INLET SAMPLE
SEPTUM
iV
1 1 \
II ,
11
I1
'l 1
I. 1
M 1
( ' '
T^
1 »
HI
^
<
SACTOR I
SATING B
OUTLET SAMPLE
^*~ SEPTUM
>• EXHAUST GAS
TO VENT
N
LOCK
FIGURE 1 - Flow Diagram of Reactor System
12
-------
THERMOCOUPLE
LOCATIONS
GAS PREHEAT
CHAMBER '
GAS INLET
EXHAUST GAS TO COOLER
POROUS PLATE
POROUS PLATE
THERMOWELL TUBE
FIGURE 2 - Details of Flow Reactor
13
-------
CAHN R-100 ELECTROBALANCE
GAS OUTLET
FURNACE
TEMPERATURE
RECORDER
PROPORTIONAL
TEMPERATURE
CONTROLLER
RECORDER
T
ROTAMETER
TIME DERIVATIVE
COMPUTER
HELIUM PURGE
TARE PAN
ULFIDE SAMPLE
ROTAMETER
TEST GAS
FIGURE 3 - Diagram of Electrobalance Reactor System
-------
suppression ranges capable of electronically taring from 10 micrograms to 10
grams. The readibility of the balance is 0.5 micrograms with five weight
ranges: 10 grams, 1 gram, 100 milligrams, 1 milligram, and 100 micrograms
full chart scale. The precision of the instrument is ±10 of the meter (or
recorder) range and ±10 of load while the accuracy is ±5xlO~4 of mass sup-
pression range for absolute weighings. The maximum weight change is 10 grams,
either increase or decrease.
The system shown in Figure 3 operates with the feed gas passing through
a rotometer to the bottom part of the reactor. Exhaust gases leave the
reactor below the glass bell housing which contains the weighing mechanism.
A continuous helium purge is run through the bell housing to keep the balance
mechanism free of the corrosive feed gases. Although most of the weighing
mechanism is gold plated to resist corrosion, some parts are subject to attack
by NO . The reactor hang down tube was enclosed in a vertically mounted tube
furnace which could easily be opened to allow access to the reactor tube be-
fore and after a run. It was controlled by a Teco TC-1000 proportional tem-
perature controller.
The cross section of the reactor hang down tube is shown in Figure 4.
In powdered sulfide tests, the reactant rested on a circular stainless steel
pan suspended by a 0.1 mm nickel wire from one side of the balance arm. An
identical pan suspended from the other side of the balance arm was used for
the tare weights. Some of the tests using sulfides impregnated on high sur-
face area support pellets were carried out using a 200 mesh stainless steel
screen suspended from the balance arm. During the course of the investigation
two reactor hang down tubes were used, one 16 mm diameter, 780 mm long ATM
Flothru mullite with ground glass joints top and bottom, and the other a 57
mm diameter, 840 ram long ATM Flothru Vycor tube with the same joints. In op-
eration a pyrex glass connector containing a porous glass plate was attached
to the bottom joint. Two chromel-alumel thermocouples were cemented with
epoxy into a hole in the side of the connector and extended up into the hang
down tube to a point just below the support pan. One thermocouple was at-
tached to a proportional controller, the other to a temperature recorder.
The hang down tube-connector assembly was filled with 40 mesh Ottawa sand to
a point just below the end of the thermocouples to aid in preheating the feed
gas. The gas was fed into the bottom of the reactor and flowed past the sul-
fide being tested before being exhausted out the top.
To measure the rate of reaction, the particular sulfide, either in pow-
der or pellet form, was placed on the weighing pan (or wire mesh pan) and the
reactor heated to the desired temperature while continuously purging the sys-
tem with helium. This was done until the weight of the sample remained con-
stant since there was usually some weight loss due to the removal of moisture
in the sample. After a constant sample weight was reached, the test gas was
introduced into the reaction tube through a rotameter and the weight change
continuously recorded as the gas-solid reaction proceeded.
The slope of the weight versus time curve represents the rate of reac-
tion and can be calculated graphically, or by use of the time derivative com-
puter which is also part of the system. Since the rate of reaction gradually
decreases with time because of diffusional resistance through the outer
reacted layer of solid, the weight change during the first hour was usually
used to determine the reaction rate for comparison purposes. This simple
15
-------
SUSPENSION WIRE
POWDERED SULFIDE
SUPPORT PAN
. 4j__ THERMOCOUPLES
OTTAWA SAND
FEED INLET
POROUS GLASS PLATE
FIGURE 4 - Electrobalance Reactor Details
16
-------
procedure was not satisfactory for comparison of the ratio of reaction for
CaS impregnated on high surface area pellets because of the large variation
of rate with time. Integrated average rates calculated for the entire run
time were used for these comparisons.
During the late stages of this research, a chemiluminescent analyzer
became available which permitted the analysis of NO down to the 10 parts per
million range. With this addition, the testing of fee metal sulfide for NO
reduction using a synthetic flue gas became feasible. x
For the tests using the synthetic flue gas, a reactor very similar to
those previously discussed was used, except that flow was down through pow-
dered sulfide rather than up. Better reproducibility was obtained with this
arrangement, probably because it eliminated some of the channeling which was
evident in the up-flow reactor. Figure 5 is a schematic flow diagram of the
apparatus used for these tests. The synthetic flue gas was made by mixing
the various gases with a gas containing 0.5 percent NO in N using calibrated
rotameters. Since the NO analyzer requires a minimum continuous flow of
sample gas to it, under the conditions of some tests, dilution was required
to provide sufficient flow to the analyzer. This was accomplished by mixing
pure N2 with the gas stream before it entered the analyzer. A metal bellows
pump was available in the system to provide a high gas recycle rate through
the reactor. At high enough recycle rates, the reactor performance approaches
that of a perfect mixed reactor and permits the simple determination of reac-
tion rates under conditions in the flow reactor.
ANALYSES
Depending on the test gas and the reaction conditions, the exit gas could
contain varying amounts of O N , NO, NO NO, SO and CO and thus the
analytical problem is quite complex. For mucn of tne preliminary work, only
a gas chromatograph was available for analysis and this put limitations on the
specific combinations of compounds that could be quantitatively determined.
Thus, for much of the work reported in this report, a test gas containing NO
in helium was used, and the reacted gas stream was analyzed for N to deter-
mine reduction of the NO. A two column system was used for most of the gas
analysis in a Varian Aerograph Model 1420 low volume thermal conductivity
chromatograph. This instrument is capable of detection in the 1-10 ppm range.
The following columns and conditions were used:
internal Column External Column
Packing 12' x 1/8" Stainless Steel 25' x 1/8" Stainless Steel
Porapak Q-S Porapak Q
Column Temp 130°C 0°C
Dector Temp 130°C 130 °C
He Flow 10 cc/min 10 cc/min
In the first (internal) column, N and NO emerge as one peak while NO and
SO are eluted separately. N ana NO are separated in the low temperature
(external) column which was enclosed in a Dewar flask containing ice. Equal
gas volumes are injected into each column using a gas tight syringe and the
chromatograph polarity switched manually at the proper time to obtain both
sets of peaks.
17
-------
1 .
u u fiT
' • ' T T
* . : : : V'
\
f\
Analyzer |
Diluent
Rotumctcr
Rotameter
— II <
M Manometer Recycle Line
X On -Oft Valve
T Control Valve fgV *
• owersioi
^.^ fo\
Q3 Recycle Pump ^^
0 Pressure Gauge
D Filter [_J>
~2. Cooling Section
V
\
)
f \
S. 4
i
3
— -
!_
^
f
\
^J
-By-Pass
^-Reactor
Housing
r " "
i
.' / •
! /
— - -- ^
Temperature
Recorder
FIGURE 5. Schematic of the Apparatus used with the Tubular Reactor for
Synthetic Flue Gas Tests
18
-------
During the late stages of the project a Thermo-Electron Model 10A Chemi-
luminescent analyzer became available and this was used to determine NO and
NO concentrations, especially when working with NO concentrations below
I,v5o0 PPM when O was also present. N and 0 concentrations were determined
using a molecular sieve 13X column in the chromatograph.
Reacted solids were analyzed for sulfate using simple barium precipita-
tion techniques.
19
-------
SECTION V
EXPERIMENTAL RESULTS
PRELIMINARY SCREENING
For the preliminary tests, approximately 20 grams of the sulfide were
placed in the reactor and then pure NO was injected at a rate of about 0.3
liters/hour while the reaction zone was being heated. Samples of the exit
gas were taken periodically and analyzed by the gas-chromatograph for NO, NO
and N . The temperature at which NO or N was first detected was noted and
the heating was continued until complete conversion (within the sensitivity
of the chromatograph) of the NO to N was obtained. The reactor was then
purged with helium while cooling and then the reacted sulfide was weighed and
when positive weight gains were noted tested for sulfate by the barium pre-
cipitation method. The results of these preliminary tests are presented in
Table 2. As can be seen from the data presented in Table 2, virtually all of
the metal sulfides tested resulted in the reduction of pure NO to N but the
temperature range in which the reaction proceeds varies widely. Actually, the
temperature at which the reaction will proceed also probably depends on a num-
ber of uncontrolled variables such as particle size (surface area), bed pack-
ing, and fluid-particle dynamics. But these data give an indication of over-
all reactivity of the various sulfides and was sufficient for the preliminary
tests.
Many of the sulfides reacted with pure NO to form at least some sulfate
but analysis for other possible reaction products was not carried out and
other oxidation products are also possible. Al S , Sb S , BiS, CuS, FeS,
MnS, HgS, MoS , K S, WS , and Tl S all lost weight during reaction. Visual
inspection showed that considerable elemental mercury was formed when HgS was
tested, but the final form of the other materials was not determined and most
of them were not considered further. A12S3 ^s unstable an<^ reacts with atmo1-
spheric water to form H S and, as a result, probably would not be practical
for use under any conditions.
Based on the success of the initial tests, several of the sulfides which
appeared to have promise were tested further to see if the reduction reaction
could be catalytically promoted. The first candidate catalysts tested were
Fe O and Cr O since these have been reported to increase the rate of reduc-
tion of NO witn carbon monoxide (3) but the catalytic effects of these
materials were negligible. However, the addition of about 5 Wt. % CoCl_ or
FeCl to CaS resulted in a pronounced (about 200°C) decrease in the tempera-
ture at which the reduction reaction proceeded although they appeared to
exhibit this effect only in the temperature range between 250 to 300°C. The
.reaction rate declined at temperatures above 350°C and the mixtures were
relatively inactive until a temperature of about 450°C was again reached.
Based on the success of these tests, other compounds were tried on a trial
and error basis with the tests being limited to materials on hand.
20
-------
TABLE 2. PRELIMINARY TESTS OF VARIOUS METAL SULFIDES FOR NO REDUCTION
Temperature
Reducing of Reaction
Agent Initiation, °C
A12S3
Sb2s3
BaS
BiS
CdS
CaS
CuS
Cu2S
FeS
PbS
MnS
HgS
MoS
SrS
K S (sulfurated potash)
ws2
ZnS
NiS
T12S
250
200
475
350
650
450
250
450
300
250
475
450
400
400
100
250
550
90
—
Temperature
for 100%
Conversion
400
550
650
600
800
650
500
700
525
550
550
650
500
650
450
500
800
360
—
Weight
Change
loss
loss
gain
loss
gain
gain
loss
gain
loss
gain
loss
loss
loss
gain
loss
loss
gain
gain
loss
Sulfate
Test
7
+
+
+
+
+
+
+
+
For these tests 20 grains of powdered sulfide were physically mixed with
20 Wt. % of the candidate catalyst using a mortar and pestle. Again, the
test gas was pure NO at a rate of 0.3 liters per hour. The results of these
tests are presented in Table 3.
As can be seen from Table 3, the addition of CoCl2, NiCl , FeCl2, and
K CrF to CaS significantly reduced the temperature at which the reduction
reacton proceeded while Fe O , Cr^, CaCl2, CuCl2, FeSO4, K ZrF ,
and PtCl had little or no effect on the reaction. K FeF ana Nicl
reduce tiie reaction temperature with BaS, SrS, ZnS, MoS2/ and FeS2. NaCl
and CaCl both appear to enhance the reaction with BaS while CuCl appears
exhibit a catalytic effect with Cu2S, MoS2 and FeS2.
A violent, highly exothermic reaction occurred when argentic (silver)
fluoride, AgF , was mixed with CaS, and so this material was not tested fur-
ther. 2
These tests showed that at higher temperatures when using some of the
promoters another peak showed up in the gas analysis using the internal Q-S
21
-------
TABLE 3. CATALYST STUDIES FOR NO REDUCTION USING VARIOUS SULFIDES
TEST GAS PURE NO
Temp, for Initial Temperature at Which Reaction
Reduction with Product Observed, °C
Unpromoted 100% NO Weight
Catalyst Sulfide Sulfide, °C NO N Conversion Change
K,FeFc BaS 475 200 350 700 gain
J D
BiS 350 450 350 — loss
CdS 650 350 350 — loss
CaS 450 — 250 400 gain
CuS 250 350 250 550 gain
FeS 300 250 200 600 gain
MoS 400 300 300 450 loss
SrS 400 200 200 500 gain
ZnS 550 500 500 700 loss
CuCl2 CdS 650 300 300 — loss
CaS 450 — — 450 gain
Cu2S 450 ~ 350
CuS 250 350 350 -- loss
FeS 300 100 100 — loss
MoS2 400 350 350 550 gain
CaCl BaS 475 200 300 600 gain
CaS 450 250 400 650 gain
FeS 300 150 250 550 no change
NiCl2 BaS 475 100 300 — gain
CdS 650 100 250 — no change
CaS 450 100 325 450 gain
CuS 250 100 400 — loss
FeS 300 50 200 550 loss
NaCl BaS 475 400 450 650 gain
CdS 650 — 600 — no change
CaS 450 300 450 700 gain
NaF BaS 475 250 300 450 gain
CdS 650 350 450 — gain
CaS 450 250 400 — gain
22
-------
TABLE 3 (continued)
Catalyst
Nal
K,FeF-+
3 D
NaCl
Fe2°3
Cr2°3
C°C12
FeCl,,
2
FeSO^
4
K,ZrF,
3 6
K3CrF6
PdCl2
Ptci2
AgF2
Sulfide
BaS
CdS
CaS
BaS
CdS
CaS
CaS
CaS
CaS
CaS
CaS
CaS
CaS
CaS
CaS
CaS
Temp, for Initial
Reduction With
llnpromoted
Sulfide, °C
475
650
450
475
650
450
450
450
450
450
450
450
450
450
450
~"~
Temperature at Which Reaction
Product Observed, °C
100* NO
NO N Conversion
250 400
500 500
250 400 650
250 350 450
150 450
250 300 550
450
450
250
255
390
400
300
450
450
__ _.. __
Weight
Change
gain
gain
gain
gain
no change
gain
—
—
—
__
_ _
-_ —
__
Porapak column. This was identified as SO . Further tests were carried out
to better define the conditions for SO formation. These data are presented
in Table 4. The same conditions as used in the catalyst tests were used in
this study.
Significantly, the reduction reaction proceeds at temperatures below the
temperature at which SO is produced and SO was not formed from BaS or CaS
under these conditions. Several preliminary runs also investigated the
reduction using a dilute NO feed stream. Helium was chosen as the diluent
for these tests because of its inertness and because it did not interfere
with the chromatographic analysis. Table 5 presents the tests using a gas
stream containing 2.5% NO in He using the same run conditions and procedures
as before.
These data are somewhat different from the results presented when 100%
NO was used as the test gas using the same catalysts. The temperature at
which N 0 and N were first observed is generally lower and the temperature
require! for 100% conversion of the NO is somewhat higher. This is a result
of using a lower attenuation on the chroma to graph for the more dilute samples.
At the higher sensitivity the NO, NO, and N were more readily detected.
£ £
23
-------
TABLE 4. PRELIMINARY TESTS FOR SO FORMATION
System
Cu_,S-NiCl_
2 2
FeS -NiCl_
2 2
CdS-NiCl2
BaS-NiCl
CaS-NiCl0
2
BaS-NaCl
CaS-NaCl
CdS-NaCl
BaS-CaCl
Temperature at Which Reaction
Observed , °C
S02 N20
550 100
450 50
450 100
100
100
250
250
—
200
Product First
N2
400
200
250
300
325
150
50
300
300
To further simulate what might be present in an actual effluent gas
stream, further runs were made with gas stream containing both O and O plus
HO using the CaS-K FeF system. Test gas compositions were: 2.5% NO, 5.0%
O and 92.5% He and 2.4% NO, 4.9% O , 2.7% HO and 90% He. Test results were
nearly identical to the tests of the dilute gas stream without O or HO al-
though with HO present there appeared to be some H S formed.
The analysis of the gas mixture containing both NO and O presented some
problems since they react to form NO and this appeared to be irreversibly
absorbed in the chromatograph column. However, the complete disappearance of
an NO peak together with the appearance of a larger N-j+O^ peak and a N.O peak
indicated that the reduction reaction was occurring. In addition, there was
a consistent weight gain of the sulfide-catalyst mixture.
TABLE 5. REDUCTION USING 2.5% NO IN He
Temperature Temperature at Which Reaction
for 100% Product First Observed Weight
System NO Conversion N20 N Change
- ;£^
CaS-K3FeF6 600 50 150 gain
CaS-K FeF +NaCl 600 100 200 gain
BaS-NaF 550 100 150 gain
BaS-K ,FeF,+NaCl 600 100 200 gain
3 O
To further characterize the solid product of reaction, a run lasting nine
days was made using calcium sulfide. Pure NO was passed through the reactor
24
-------
while it was maintained at temperatures between 450 and 700°C. After nine
days, N was still being produced but at a very low rate thought to be
limited by diffusion through the outer reacted layer. The solid reaction
product analyzed 87.2% CaSO
DISCUSSION OF PRELIMINARY RESULTS
The preliminary results were very interesting and encouraging. Nineteen
sulfides were tested for NO reduction and all gave positive results. Of
these, BaS, CaS, CdS, Cu2S, NiS, PbS, SrS, and ZnS all gained weight during
reaction indicating that undesirable side reactions producing gaseous sulfur
compounds may not be occurring. The alkaline earth sulfides BaS, CaS and SrS
looked particularly attractive. Of these, CaS probably would be preferable
because of availability and cost. In addition, CaS can easily be regener-
ated from the sulfate by reduction with CO, H or Coke (6.7).
The temperature required for the reduction reaction to proceed was
lowered considerably by mixing certain materials with the powdered sulfide.
This effect may or may not be catalytic in nature and the mechanism is not
known. It is possible that the promoter takes part in the reaction and is
chemically altered and thus not a true catalyst. The most active materials
found were metal chlorides and fluorides with K FeF , NiCl , NaF, CoCl , and
FeCl2 being the most active. It is interesting to note that certain metal
chlorides form metal complexes (FeNOCl , for example) with NO.
The nine day run with CaS indicated that diffusion through the outer
layer probably controls the rate of reaction after the sulfide on the outer
surface has reacted. This indicates that a fluidized bed of very fine par-
ticles or sulfide impregnated on a high surface area support may be necessary
to obtain high rates of reaction and efficient use of the reactant.
Oxidation by free oxygen may also present problems. In the preliminary
runs, only one test was made using a gas stream containing free oxygen.
Under this condition it appeared that the reduction of NO was occurring but
it could not be determined whether or not O was also contributing to the
oxidation of the sulfide.
Hence, subsequent tests were made of sulfides impregnated on high sur-
face area carriers and an investigation was made of the relative rates of
reaction of NO and O .
PRELIMINARY TESTS OF CaS IMPREGNATED ON SUPPORTS
Because of the apparent diffusional limitations through the outer reacted
sulfate layer, a study was made of the use of sulfide dispersed on high sur-
face area solid supports to see if this problem could be reduced. The prop-
erties of the support materials that were tested are shown in Table 6. These
included a high surface area silica-alumina, a molecular sieve developed for
SO adsorption, and two aluminas, one with 6% silica and the other with low
impurity levels. These were impregnated with calcium sulfide in the follow-
ing manner: The support material was first dried in a muffle furnace at
about 400°C for 24 hours and allowed to cool in a desiccator and weighed.
The dry pellets were then impregnated with Ca(NO3>2 by dissolving appropriate
amounts of the salt in distilled water and pouring this solution over the dry
25
-------
pellets. The pellets were left in the solution for about 24 hours to insure
saturation before the excess solution was filtered from the pellets. They
were then dried in air and calcined at about 400°C for 24 hours to convert
the nitrate to the oxide, CaO. The pellets were cooled and weighed again to
determine the amount of CaO that had been deposited on the pellets. The
oxide was then converted to the sulfide by placing the pellets in the reactot
and passing a mixture of 20% H S in H over them while the reactor was heated
to 500°C. The course of the reaction was followed by periodically analyzing
the exit gas stream for H S and HO. Although in most cases it appeared that
conversion was complete at about 400°C, the sulfiding reaction was continued
up to 500°C to insure complete conversion of the oxide to the sulfide. This
procedure is commonly used for forming the sulfide in hydrotreating catalysts
which are normally made via an oxide intermediate. In some cases, a catalyst
was also dissolved in the Ca(NO ) solution.
TABLE 6. PHYSICAL PROPERTIES OF SUPPORT MATERIAL
Material
Density (gr/cc)
2
Surface Area (M /gr)
Pore Volume (cc/gr)
Pore Diameter (A)
Si02(%)
A12°3(%)
Na20<%)
Physical form
Nalco
1290- A
0.35
270
1.22
180
74.7
23.0
0.05
1/8"
extrusions
Harshaw
AL-1602-T
.83
210-240
0.48
91-80
6.0
91.0
— —
1/8"
pellets
Alcoa
H-51
.95
350
0.50
60
—
—
—
8-14
mesh
Linde
MS-TM- 0-1114
—
—
0.53
Synthetic cry-
stalline metal
aluminum sili-
cate containing
sodium.
Material devel-
oped for SO
adsorption
1/16"
extrusions
The NO reduction tests were carried out without removing the pellets
from the reactor for weighing. Pellet sulfide compositions were assumed to
be a result of complete conversion of the CaO to CaS.
The following summarizes the results of the preliminary runs using pel-
lets impregnated with calcium sulfide.
Run 1
7.2 wt. % CaS (only) on Harshaw A1-1602-T.
standard liters per hour.
Test gas: Pure NO, 0.3
Remarks—
Some NO was formed at room temperature and the amount increased up to
about 180°C.
NO.
At 150°C the exit gas analyzed about 10% N_0, 10% N , and 80%
Some N- was formed at the lower temperature and the amount or N in exit
gas increased at temperatures above 300°C but conversion was not complete
26
-------
until a temperature of about 400°C was reached. At 400°C there was no N O
present in the exit gas. At constant temperature, the rate of reaction
decreased rapidly with time indicating diffusional limitations.
Run 2
7.2 wt. % CaS+2.5% NiCl2 on Harshaw A1-1602T. Test gas: Pure NO,
0.3 standard liters per hour.
Remarks—
At 60°C there was nearly complete conversion of the NO to NO. The main
product was NO up to about 250°C where the reaction shifted, giving lesser
amounts of NO and larger amounts of N . At 300°C the exit gas was 100% N .
The run was continued holding the reactor constant at 375°C and there was
complete conversion of the NO to N for about 7 hours. After NO breakthrough
the proportion of NO in the exit gas steadily increased.
Pellet weights: 50.00 grams original pellets
52.53 grams after impregnation/calcination
53.90 grains at end of run
Qualitative analysis of the spent pellets showed that considerable sul-
fate had been formed and that chloride was still present in the pellets indi-
cating that the nickel was still in the chloride form.
Run 3
4.5 wt. % CaS+1.5% NiCl on Alcoa H-51. Test gas: Pure NO, 0.3 stand-
ard liters per hour.
Remarks—
Using this material the reaction started at about 65°C when the exit
gas analyzed about 35% NO, 65% N . The proportion of NO increased up to
about 100°C when the exit gas was about 85% NO, 15% N with no detectable
NO. At 270°C the composition was 40% NO, 45% N and 15% NO. At 380°C it
was 95% N and 5% NO. The proportion of NO steadily increased as the temper-
ature was held constant at about 390°C for the remainder of the 8-hour run
even though the gas feed rate was dropped from 0.3 to .07 standard liters
per hour.
Pellet weights: 50.0 grains original pellets
51.8 grams after impregnation/calcination
53.6 grains at end of run
Run 4
5.3 wt. % CaS+1.8% NiCl on Linde TM-0-1114. Test gas: Pure NO, 0.3
standard liters per hour.
Remarks—
Using this material it appeared that a large amount of NO was physi-
cally adsorbed by the pellets at room temperature when the NO feed was first
introduced since the analysis showed mostly He (purge) in the exit gas for
the first 1/2 to 3/4 hour. After about one hour of operation (during which
27
-------
time the reactor was heated to 100°C) the exit gas appeared to be 100% N but
as the reactor was further heated, NO appeared and the amount increased with
increasing temperature. At 190°C the exit gas analyzed 26% NO, 74% N ; at
250°C, 50% NO and 50% N , and at 285°C, about 3% NO and 97% N . The temper-
ature was then held constant at about 300°C as the test was continued and the
exit gas analyzed 100% N for the next four hours. Then, over the next hour
period, the composition changed to about 10% NO, 45% N , and 45% NO. The
temperature was then increased to 500°C where conversion was again to 100% N
at an NO feed rate of 0.07 standard liters per hour.
Pellet weights: 50.00 grams original pellets
52.18 grams after impregnation/calcination
54.10 grams after NO reaction
Run 5
8.1 wt. % CaS+2.7% NiCl on Nalco 2910-A. Test gas: Pure NO, 0.15
standard liters per hour.
Remarks—
For this run only 30. 9 grams of pellets could be loaded in the reactor
because of their low bulk density. Using these pellets there was little
reaction at the lower temperatures. At 110°C the exit gas analyzed 17% NO,
3% N , and 80% NO. At 280°C it was 56% NO, 27% N , and 17% NO. Complete
conversion to N was attained at 475°C. A temperature of 510°C was then
maintained for the next three hours of the run. Analysis of the exit gas
showed that there was complete conversion to N at this high temperature but
it also showed that a large mount of SO was also being evolved. The tem-
perature was then reduced to about 425°C where the exit gas contained no SO ,
but there was about 30% NO in the exit gas at this temperature.
Pellet weights: 29.8 grams original pellets
30. 9 grams after impregnation/calcination
31. 2 grams after NO reaction
Subsequent tests were made using a test gas containing 2.5% NO in helium to
determine the effects of space velocity on the reduction of NO.
Run 6
7.2 wt. % CaS+2.5% NiCl., on A1-1602-T.
A,
Remarks—
For this test the reactor was purged with He as it was heated to 230°C
and the reactor was held at this temperature and the pellets tested with the
2.5% NO - Helium test gas with the following results:
28
-------
°C Time on Stream
Temp. Hours
230 0-5
5-6
6-7
7-8
8-10
325 10-12
12-13
13-15
440 15-17
Gas rate -
Standard liter/hr.
2.2
6.5
12.2
22.7
2.2
2.2
6.5
13.2
13.2
Exit Gas Composition (He free)
100% N
100% N_
2
35% N , 65% NO, trace NO
20% N0, 80% NO, trace N O
2 2
40% N , 40% NO, 20% NO
100% N
2
81% N , 9% NO, 10% NO
SO in exit gas
large amounts of SO in exit
gas - rest 100% N
Run 7
12% MoS on A1-1602-T
Remarks—
These pellets were made by impregnating the pellets with MoO dissolved
in NH OH solution, drying, calcining and converting the oxide to the sulfide
in the usual manner. They were treated with the 2.5% NO in helium gas with
the following results:
°C Time on Stream
Temp. Hours
167
200-325
325
0-4
4-13
13-16
Gas rate -
Standard liter/hr
2.2
2.2
6.5
Exit Gas Composition (He free)
65% N , 35% NO
100% N
large SO peak,
100% N
otherwise
325
16-20
2.2
large SO peak, otherwise
100% N
Pellet weights:
50.00 grams original pellets
56.16 grams after impregnation/calcination
57.85 grams after NO reaction
Run 8
7.3% ThS on A1-1602-T
Remarks—
These pellets were made by impregnating the alumina with a solution of
thorium nitrate, drying, calcining and converting the oxide to the sulfide
29
-------
in the usual manner. They were tested using the test gas containing 2.5% NO
with the following results:
°C Time on Stream
Temp. Hours
90-510 0-14
510 14-17
Gas Rate -
Standard liter/hr.
2.2
22.7
Exit Gas Composition (He free)
100% N , all temperatures
100% N , trace NO, large
S02 peak
(An SO peak appeared the
last 3 hours of the run.)
Pellet weights:
50.00 grams original pellets
53.26 grams after impregnation/calcination
53.45 grams after NO reaction
Several important conclusions can be made from these tests. First, it
appears that a very active form of the sulfide can be made by dispersing it
and a catalyst on a high surface area support. If it were desirable to de-
velop a process based on sulfide pellets, the selection of the support mater-
ial would be very important since the four materials evaluated behaved very
differently in the testing. The performance of the two aluminas and the
molecular sieve pellets were satisfactory from a standpoint of initial activ-
ity but there appeared to be significant diffusional resistance after only a
few hours of operation, especially with the molecular sieve and H-51 alumina
pellets. The performance of the high-silica pellets was unsatisfactory be-
cause of a low activity and because of the side reaction which consumed sul-
fur. The 6% SiO pellets (A1-1602-T) also produced SO during the latter
part of the runs.
The mechanism of the SO formation is not known but there are several
possibilities, among them:
CaSO + SiO •* SO + CaSiO + 0.50
4 ^ ^ J £,
CaS
+ SiO + 3NO -»• SO
ff
+ CaSiO + 1.5N
J f»
(1)
(2)
Reaction (1) has been studied as a means of desulfuring gypsum but apparently
the reaction requires temperatures on the order of 1250°C when using bulk
quantities of gypsum and silica sand (8). Since there is considerable silica
in both the A1-1602-T and 1290-A pellets, both reactions are distinct possi-
bilities. This follows since SO was only observed after a run had proceeded
for some time when there was considerable CaSO present and at higher temper-
atures.
The tests made with MoS and ThS dispersed on alumina pellets were in-
teresting because of their high activity and because analysis showed that
very little sulfate was formed. However, with these materials there was
also a large amount of SO formed late in the run and so these materials
were not investigated further.
30
-------
At temperatures below about 250°C and at higher space velocities, there
was considerably NO formation. This may be acceptable, however, because N O
is less of a concern than NO and NO
£
The low capacity of the impregnated pellets is disturbing and thought to
be due to diffusional limitations. There are at least two possible mecha-
nisms for the rapid decline in activity with use. If the reactions proceed
throughout the pellet, there still may be diffusional limitation caused by
diffusion limitations through the CaSO layer dispersed throughout the pel-
let. On the other hand, if the reaction proceeds inwardly from the outside
of the pellet, severe pore diffusion limitations may result.
Figure 6 presents the results of a series of 3 runs to show typical
break-through curves. In each of the 3 runs, 13 grams of pellets from the
same batch (7.2% CaS+2.5% NiCl on AL-1602-T) were evaluated using a test gas
containing 2.5% NO in He with a gas flow of 5 liters/hour. As can be seen at
300°C there is NO in the exit gas from the beginning of the run while about
4.5 hours were required for break-through at 400°C.
The upper curve in Figure 6 presents results of a run using material
from the same batch but crushed and screened to 14 mesh. This was done in an
attempt to test for pore diffusion; however, the results of this run are
poorer than with the whole pellets. This may have been due to poor contact
because of channeling in the bed of smaller particles or, a more likely pos-
sibility is that there was air oxidation of the pellets during crushing and
screening. In comparing the results of these tests with Run #6 that was
previously described, it can be seen that the performance was not as good as
in Run #6. This further indicates air oxidation of the sulfide since these
pellets were exposed to air while those in Run #6 were not.
Subsequent tests were aimed at determining the rate of reaction of CaS
impregnated on the different supports and comparing the rate of reaction of
some of the sulfides with NO and 0 . The electrobalance was used for these
studies.
ELECTROBALANCE STUDIES
The pellets for this study were prepared as previously described and
the pellets were stored under a nitrogen blanket until needed. For these
tests, only CaS was impregnated on the pellets since it appeared that the
addition of a catalyst such as NiCl also resulted in SO being formed.
A prepared pellet was tested by placing it on the weighing pan of the
electrobalance and the hang down tube was purged with pure helium while the
tube was heated to the desired temperature. The purge was continued until
the weight remained constant (there was usually a weight loss due to mois-
ture loss). Then a 2.5% NO, 97.5% He mixture was passed through the tube at
a rate of about 11 cc/min. For all of the tests except one only, one pellet
was placed on the weighing tray for testing. In the other test, 5 pellets
were used. Table 7 gives the reaction conditions for the various tests.
The results of the reaction rate determination are conveniently shown
on plots which indicates both the pellet weight and the rate of reaction as
a function of time. The rates of reaction (which is the slope of the weight
31
-------
100
U)
N)
en
O
Z
80
CRUSHED PELLETS
60
40
20
Pellet Composition: AL-1602-T
with 7.2% CaS, 2.5% NiCl2
Gas rate: 5 LITERS/HR.
2.5% NO in He
I
8
10
12
Time, hours
FIGURE 6 - Break Through Curves for CaS Pellets Promoted with NiCl
-------
versus time curve at any point) were determined by evaluating the slopes of
the weight curve by a least squares technique over small intervals.
TABLE 7. REACTION CONDITIONS
Run
No.
1
2
3
4
5
6
7
8
9
10**
11
12
13
Type
of
Support
Harshaw
1602-T
Harshaw
Harshaw
Harshaw
Harshaw
Harshaw
Harshaw
Harshaw
Harshaw
Harshaw
Linde
TM- 0-111 4
Linde
Linde
Temp.
°C
390
410
437
438
468
493
442
440
444
444
392
410
438
Test Gas Helium Purge
Flow Rate Thru Bell Jar Total
CaS (Std.cm3/ (Std.cm3/ Run
(Wt. %) sec.) sec.) Time (hr)
8.6
8.6
8.6
8.6
8.6
8.6
8.6
.9*
.9*
.9*
6.6
11.4
6.6
.18
.15
.16
.18
.20
.18
.16
3.48
.12
.12
.14
.14
.17
.046
.045
.045
.045
.032
.045
.045
.052
.052
.063
.043
.043
.043
7
9-1/2
8
9
5
4
18
4
3-1/2
4-1/2
12
4
18-1/2
No. of
Pellets
1
1
1
1
1
5
5
T_
1
1
1 >
1
1
* Pellets were pre-oxidized by contact with air
**Feed gas composition was 2.5% O , 97.5% He
The time derivatives are shown on the plots only as points rather than
curves because of the scatter in the data. This scatter is typical of deriv-
ative data.
The results of the runs are presented in Figures 7 through 18. Some
of the data exhibit periods of increasing rate of reaction several hours
into the run after a period of steadily decreasing rate. These are all
thought to be due to changing controlling resistances and non-isothermal
effects caused by the highly exothermic reaction. The two lower temper-
ature runs, Figures 7 and 8, are examples of this.
Figures 9 and 10 show the results of runs carried out at nearly the same
temperature to determine reproducibility. The results of these two runs are
in fair agreement as can be seen by comparing the rates at any given time.
For the first three hours the rates are approximately constant at about
0.45xlO~4 moles/hour per gram and then the rate gradually drops to zero after
about nine hours. There are many mechanism that can explain this behavior.
33
-------
u>
31.5
31.3
31.1
£
4*
V
w
04
30.9
30.7
0°.
o
°cP
o o
o ,
Harshav
390°C
2.5*NO, 97.5*He
.18 std cm3/sec
cP
8
Time, hr
10
Figure 7. Reaction data for a Harshaw pellet at 390°C
1.0
.8
I
o
(0
OJ
H
i
-P
OJ
0)
Pi
X
4)
.2
-------
U)
Ul
32.
32.2
DO
3
32.0
$31.8
31.6
•15 std cm /sec
o o
°°0
cP
°°o0o0
8
10
Time , hr
A A**»^ ) ££4.
Figure 8. Reaction data for a Harshaw pellet at UlO C
1.0
.8
0)
o
Vi
to
0)
r4
o
o>
r-{
•-(
V
.1*
A)
13
.2
-------
u>
32.8 •
32.6
4)
32. U
32.2
32.0
Harshav
U37°C
2.5*NO, 97.5*He
^
.16 std on /sec
8
10
Time, hr
Figure &. Reaction data for a Harshav pellet at *O7 C
1.0
.8
fi
bO
w
OJ
X
«
.2
-------
Ul
30.1
29.9
3
3
V
29.7
29.5
o o
o o
QO o
°oo
, 97.5*He
•18 stfl cm /sec
.8
,
I
on
0)
.H
i
&
,]*
.2
K
4)
u
K
2 U 6 8 10
Time, hr
Figure 10. Reaction data for a Harshav pellet at ^38 C
-------
However, it may be an indication that the reaction is occurring throughout
the pellet and the rate remains high as long as easily accessible CaS re-
mains on the surface. When the CaS on the surface is reacted, the reaction
may be controlled by a slower diffusion of the NO through reacted layers of
sulfate. If pore diffusion were controlling for this situation, there would
be a very rapid decline in activity as the material near the outer surface
was consumed.
At the higher temperatures investigated (468 and 493°C) , shown in Fig-
ures 11 and 12, pore diffusion may be the controlling resistance since for
these runs there is a period of nearly constant reaction rate followed by a
very rapid decline in activity over a two hour period.
Figure 13 presents the results of a run made at 442°C with 5 pellets on
the pan in a pile (some on top of others) in a random manner. The weight
versus time curve exhibits a classical sigmoid shape for this case. This is
thought to be due to non-isothermal temperature effects in the pile caused
by the highly exothermic reaction. The average rate of reaction for this
case was about an order of magnitude less than for the single pellet runs.
This is probably due to stagnant gas pockets within the pile since gas flow
was around the pan rather than through the pile. The single pellet data are
probably more representative of the rate that would be obtained in a packed
bed reactor where gas flow would be through a bed of pellets.
Figures 14 and 15 are a comparison of the rate of reaction of the Har-
shaw pellets with NO and 0 . As can.be seen, the initial rate of reaction
with 0 was very high (about 2.4x10 moles/hr. gram) being about 6 times
the rate observed for NO. However, the reaction quickly slowed and after the
first hour was at a rate comparable to that with NO.
The last three pellet tests on the microbalance used Linde Molecular
Sieves TM-0-1114. These data are presented in Figures 16 through 18 for tem-
peratures of 392, 410, and 438°C respectively. All of the runs made with the
molecular sieve pellets exhibited a very high initial rate of reaction (up to
3 times the rate observed with the Harshaw pellets) but there was a very
rapid decline in rate. As previously discussed, this is typical of a system
in which pore diffusion is important and activity rapidly declines as the CaS
near the outer surface reacts.
Because of difficulties in comparing the runs made at different temper-
atures, average rates for each run were computed. Table 8 presents the inte-
grated average rates for each run. Though instantaneous rates vary without
much pattern, the average rates consistently increase with temperature as
would be expected.
For the single Harshaw pellets the rate increased with temperatures from
0.25x10* to 0. 45x10 moles of CaSO formed per hour per gram of initial
pellet weight for a temperature increase from 390 to 493°C. Kinetically,
this probably means that the overall rate of reaction is being controlled by
diffusion since, if the reaction were controlling, the rate should roughly
double for a 10°C temperature increase.
The average rate of reaction for the run with the test gas containing
s tt
38
2.5% O instead of NO was about twice that for NO on a CaSO basis; however,
-------
LJ
VO
3U.2
i
•H
a
3U.O
Harshaw
2.5*NO, 97.5*He
• 20 std cm /sec
.8
I
a
93
0)
ft
ttJ
33.8
O O
8
10
Time, hr
Figure 11. Reaction data for a Harshaw pellet at U68 C
.2
x
v
K
-------
32.2
32.0
£
•H
O
H 31.8
31.6
Harshav
l*93°C
2.5*NO, 97.5*He
.18 std cm /sec
2 U 6 8 10
Time, hr
Figure 12. Reaction data for a Harshav pellet at U93 C
•a
I
o
.6 o
o
to
o
a
•p
0)
rH
H
Pi
00
IS
K
-------
263.9
263.1*
- 262.9
(X,
262. U
261.9
Harshaw
1*1»2°C
Q
.16 std cm /sec
00
16
20
U 8 12
Time, hr
Figure 13. Reaction data for five Harshaw pellets at UU2°C
1.0
.8
0)
p«
0)
(1)
iH
O
a
IA
O
H
X
«
-------
27.5
21
fi
•H
0)
2 27-3
0)
(L,
27-2
, 97.5*He
•12 std cm /sec
12345
Time, hr
Figure 14. Reaction data for a Harshaw pellet using oxygen
2.0
•o
o
-------
25.6
25-55
OJ
•P
.C
0)
25.5
25.^5
Harshav
.12 std cm /sec
.8
•CJ
i
CO
(II
O
tn
o
0)
0)
O
X
0)
"rt
cc
O O
1 2 • 3 i* 5
Time, hr
Figure 15. Reaction data for a Harshaw pellet at low flow rates
-------
25.50
25.25
•Jj 25.00
Al
2U.75
.8
I
.6
&
(0
0)
H
O
0)
ts
.2
2 1*. & 8 10
Time, hr
Figure 16. Reaction data for a Linde sieve at 392°C
-------
Linde
Ul
26.50
. 26.25
ti
o»
cu
26.00
25.75
O « O
, 97.5*He
.lU std cm/sec
.8
I
o
0!
0)
rH
O
a
r-l
(II
.2
o
H
0)
•8
K
Time, hr
Figure 17. Reaction data for a Linde sieve at UlO C
-------
22.0
t» 21.5
a\
21.0
20.5
Linde
U38°C
2.5*10, 9
.17 std cm /sec
.00
CO °QO 0000°
I
8
12
Time, hr
20
Figure 18. Reaction data for a Linde sieve at 1|38WC
1,6
11.2
0)
1
o
00
i
V
6
04
.8
a>
•P
-------
on a per mole of 0 basis, the rates are comparable since the rate of reac-
tion is twice that for O per mole of CaSO .
2 4
TABLE 8. AVERAGE RATES OF REACTION VS. TEMPERATURE
Run
No.
1
2
3
4
5
6
7*
8
9**
10
11
12
Type
of
Support
Ha r shaw
Harshaw
Harshaw
Harshaw
Harshaw
Harshaw
Harshaw
Harshaw
Harshaw
Linde
Linde
Linde
Rate x 104
Temp. mol CaSO, formed
°C
390
410
437
438
468
493
442
444
444
392
410
438
(hr) (g "pellet)
.25
.25
.34
.31
.40
.45
.05
.22
.41
.32
.64
.47
mol NO removed
(hr)(g pellet)
1.
1.
1.
1.
1.
1.
.
.
.
1.
2.
1.
00
00
36
24
60
80
20
90
92***
28
56
88
Time
Range Time
.9
.9
.9
.9
.9
.9
1.4
.7
.3
1.0
.5
.6
to
to
to
to
to
to
to
to
to
to
to
to
6.9
9.5
8.1
9.1
5.3
3.9
17.8
3.3
4.3
12.2
4.3
18.6
* 5 pellets
** pellets partially oxidized before run for runs 9 and 10 2.5% 0 ,
97.5% He feed
***Rate based on moles of O removed
The average rate of reaction for the Linde Molecular Sieve was higher
than for the Harshaw pellets at comparable temperatures with the maximum
average rate being 0.64xlO~ moles CaSO formed per hour per gram of initial
pellet weight at 410°C. This rate was higher than the rate obtained at
438°C. This inconsistency probably resulted due to deactivation of the
molecular sieve pellets because of oxidation in air. If the pellets were
exposed to air for even short periods of time, they became very hot and
analysis showed almost complete oxidation to CaS04>
Placing pellet on the balance pan required a finite but variable time
exposure to air. For comparison, the Harshaw Pellets became about 50%
oxidized after being exposed to air for about 17 days at room temperature.
All of the pellets underwent a color change upon reaction. The Harshaw
pellets were initially light blue and became light brown or beige when im-
pregnated with CaS. Upon reaction with NO they became bright white. The
molecular sieve pellets were initially brown and turned black when impreg-
nated with CaS. After reaction they again became brown.
The discovery that the CaS impregnated on high surface area pellets are
rapidly oxidized by atmospheric air at ambient temperatures was disturbing.
47
-------
It had been assumed that this would not be a problem, especially at lower
temperatures. This assumption was based on a study by the U.S. Bureau of
Mines of the reduction of CaSO to CaS using gypsum-coke pellets in a rotary
kiln which showed that CaS was not appreciably oxidized when the pellets
were discharged from the kiln at 700-800°C (9).
Subsequent tests were carried out to compare the relative rate of reac-
tion of various sulfide-catalyst combinations with NO and O . The electro-
balance was also used for these studies.
RATE OF REACTION OF METAL SULFIDES WITH NO AND O
For these tests, 0.8 grams of the powdered sulfide were distributed
evenly on the weighing pan of the balance, and the weight change noted when
a test gas mixture of either 2.5% NO or O in helium was passed through the
hang down tube at a rate of about 13 cc/min. Temperatures of 300, 400 and
500°C were investigated with a fresh sample of sulfide being used for each
temperature and each gas mixture. The reactor was purged with pure helium
while heating the reactor to the desired temperature and until the weight
remained constant (due to moisture being removed from the sample). Table 9
is a summary of the rate of weight gain or loss for the various sulfides
listed. These same data converted to rate of formation of sulfate per minute
per initial weight of the sulfide are shown in Table 10. In these tables,
the negative sign indicates a weight loss when the test gas containing NO or
O is passed through the reactor tube while the "d" indicates a weight loss
while purging with pure He at that temperature indicating a decomposition.
As can be seen from the data, the metal sulfides react faster with oxy-
gen in every case except one. At 300°C the rate of reaction of SrS was
slightly faster with NO than with O . BaS, FeS and CaS all lost weight when
NO was passed through the reactor at 300°C indicating the possibility of un-
desirable side reactions taking place. Analysis of the dilute exit gas
stream from these runs was impossible so these possible side reactions were
not investigated further. BaS, FeS, SrS, CdS, CaS and sulfurated potash
(K S) all reacted both with NO and O . MnS, T1S, CuS, and ItoS all decom-
posed in the helium purge. WS lost weight when reacting with Doth 0 and
NO. ZnS did not react with NO at 400 and 500°C but did react with O 7 This
is in line with the preliminary tests which showed that NO does not react
with ZnS at temperatures below about 550°C.
The sulfurated potash (K S plus higher sulfides) exhibited a very rapid
weight gain but the rate of formation of the sulfate could not be calculated
because the exact composition of the sulfurated potash was not known. Dur-
ing the test of the sulfurated potash at 500°C with NO, there was a large
(0.3 gram) loss in weight of the stainless steel weighing pan and so a 500°C
run using O was not made.
The fact that the rate of weight changes when the sulfides react with
O is faster than with NO is slightly misleading since it takes twice as
much NO as O on a molar basis (assuming sulfate is formed) to produce the
same weight change. Table 11 presents the same data based on the moles of
NO or 0 reacting, assuming the metal sulfide is the reaction product. As
can be seen from Table 11, on this basis, the rate of reaction of NO is
greater than O for SrS at 300 and 400°C and BaS and CdS at 400°C. Also,
48
-------
significantly, there is little difference between the rates of reaction for
NO and O for SrS at all temperatures.
TABLE 9. A SUMMARY OF THE RATE OF WEIGHT GAIN OR LOSS FOR THE VARIOUS
SULFIDES REACTING WITH NO AND O
Sul fides
BaS
FeS
ZnS
SrS
CuS
CdS
Cu^S
2
PbS
ws,.
2
MoS2
K2S (Sul fur a ted
Potash)
T1S
CaS
MnS
300
NO
-2.08
-2.27
1.50
2.22
d
.98
2.53
0
0
38.8
d
- .49
d
Rate x i
°C
°2
0
-20.80
1.77
2.15
d
3.80
8.26
-3.08
185.00
0
, ^3 grams of weight change
L(J t
minute
400 °C
NO O
4.09 5.3
1.172 27.58
0 4.53
2.73 5.29
d d
3.62 5.50
3.52 31.70
-8.33 -22.90
d
79.8 88.70
d
2.08 6.50
d
500
NO
7.22
70.9
0
3.68
d
2.43
0
1.59
-6.54
d
127.00
d
3.85
d
°C
°2
182.00
144.0
7.97
7.81
d
27.6
8.07
108 . 00
-25.20
32.40
NOTES: - represents a weight loss
d indicates weight loss while purging the reactor tube with pure He
The next series of experiments investigated the relative rates and reac-
tions of CaS with NO and O when it was mixed with various promoters and
tests of the promoters alone. The results of the tests of NaF, NaCl, K FeF
and K FeF +NaCl mixed with CaS are shown in Table 12. For these tests 20 wt
% promoter was mixed with CaS and runs conducted as in the previous runs
using test gases containing 2.5% NO or 0^.
As can be seen from Table 12, the presence of NaF or K^FeF has a pro-
nounced effect on the rate of reaction of CaS with both NO and «2 but there
is a greater increase in the rate of reaction of NO, especially at the lower
temperature of 200 and 300°C. This is very significant because it indicates
that there is a possibility that the sulfide can be used to reduce NO in a
gas stream also containing O , without excessive consumption of the solid by
reaction with O . It also probably indicates that the mechanism of the reac-
tion of NO with2the sulfide is different than the mechanism for the reaction
with 02.
49
-------
TABLE 10. REACTION RATES OF NO AND O WITH VARIOUS METAL SULFIDES
(wt. of MeSO formed)
Rate x 10
(minute)(initial wt. of MeS)
Sul fides
BaS
FeS
ZnS
SrS
CuS
CdS
Cu2S
PbS
ws2
MoS
K S (Sulfurated
Potash)
T1S
CaS
MnS
300°C 400°C 500°
NO O NO 0 NO
0.00 18.60 24.00 32.60
3.38 85.00 207.60
4.80 5.90 0.00 14.00 0.00
7.92 7.65 9.81 18.70 13.20
d d d d d
3.98 15.50 14.70 22.80 9.79
0.00
14.80 50.30 20.50 191.00 9.18
0.00
0.00 d d d
+ + + + d
d d d d d
0.00 5.58 17.70 10.90
d d d d d
C
°2
874.00
428.00
25.10
27.70
d
116.00
37.00
648.00
—
d
d
d
86.50
d
TABLE 11. THE RELATIVE RATES AT WHICH NO AND O REACT WITH THE
VARIOUS SULFIDES
Rate x 10
Sulfides
moles of NO reacted
moles of O reacted
(minute)(initial grams of MeS) (minute)(initial grams of MeS)
300 °C
NO
400°C
NO
500°C
NO
BaS
FeS
ZnS
SrS
CdS
Cu2S
PbS
CaS
—
—
1.19
1.73
.76
1.95
—
0
—
.73
.83
1.49
3.32
0
3.19
.89
0
2.14
2.82
2.70
1.64
2.06
11.20
1.73
2.04
2.19
12.60
2.56
5.59
54.80
0
2.88
1.88
0
1.21
3.20
74.90
56.40
3.11
3.02
11.10
3.32
42.70
5.71
50
-------
TABLE 12. RATE OF REACTION OF CaS AT VARIOUS TEMPERATURES
USING DIFFERENT PROMOTERS
Promoter
NaF
NaCl
K3FeF6
K.FeF +
3 o
NaCl
no
promoter
n-,^ v In7 gram moles of
(minute) (gram
200°C 300°C 400
NO O NO 0 NO
3.35 0 1.93 1.02 5.99
0 0 2.67 0 1.00
2.99 0 3.73 1.39 10.80
2.68 0 3.58 1.66 5.99
1.64
gas reacted
initial CaS)
°C
°2
18.5
2.39
21.60
47.3
2.56
500
NO
43.5
2.69
25.9
4.63
3.20
°C
°2
73.9
21.0
24.2
56.7
5.71
To further test these promoters, runs were made using the pure material
on the electrobalance to see if they reacted with either NO or O . The
results of these experiments are shown in Table 13. As can be seen, the NO
reacts with K FeF at all temperatures investigated while it does not react
with O except at 500°C. NaF does not react with NO and only reacts with O
at 500*C. NaCl does not react with either gas at temperatures of 400°C or
below. There appears to be synergetic effect on the reaction of NO with
K.FeF and NaCl mixtures at 500°C.
3 6
TABLE 13. THE WEIGHT CHANGE WITH TIME OF VARIOUS PROMOTERS REACTING
WITH NO OR 0.,
Reaction rate, (Milligrams of Weight Change/Min) x 10
Promoter
NaF
NaCl
K,FeF_
3 6
NaCl
NiCl2
FeCl2
coci2
Fe2°3
200°C 300°C
NO O NO
ooo
ooo
3.4 0 22.9
3.0 0 1.0
0
0
1.5
2.9
0,
2
0
0
0
0
0
0
0
0
NO
0
0
3.
1.
0
0
-12.
0
400°C
o0
2
0
0
0 0
8 0
0
0
9 0
0
500°C
NO 0_
2
0 11.0
7.3 113.9
42.2 45.5
606.8 29.4
51
-------
From these tests the K3FeF should probably be eliminated from further
consideration for use as a promoter. Not only does it react with both NO
and O but it undergoes a color change from white to reddish brown as it is
heatea in helium indicating that it is unstable at higher temperatures.
Another series of tests (Table 14) investigated NaF, NiCl, FeCl , CoCl
and Fe O mixed with FeS, BaS, SrS and CaS in the reduction of 2.5% NO in
helium at 300 and 400°C. As can be seen the combinations giving the highest
reaction rates were FeS-FeCl at 300°C, BaS-FeCl at 400°C and CaS-Fe O at
400°C. All combinations were investigated further using a synthetic flue gas
in the flow reactor system.
TABLE 14. THE RATE AT WHICH NO REACTS WITH THE VARIOUS METAL
SULFIDE CHEMICAL PROMOTER MIXTURES
Metal
Sulfide
FeS
BaS
SrS
CaS
FeS
BaS
SrS
CaS
Rate x
Pure
Metal
Sulfide
0
0
2.16
0
1.11
3.99
2.68
2.05
7 Moles NO Reacted
(Minute)
NaF
1.64
2.70
2.85
2.41
19.86
7.36
1.00
7.49
(Initial)
NiCl2
300°C
0
2.71
1.05
0
400°C
26.36
31.75
0
5.36
Gram Sulfide)
Promoter
FeCl CoCl2
58.63 1.33
.58
0 0
1.31 .87
1.69
39.44
0 0
1.95 2.24
Fe2°3
0
2.19
2. IB
1.00
4.19
0
6.98
32.00
Reduction of NO In a Synthetic Flue Gas
^^^^^"•^^^ X ~""" "
The reduction of NO in a synthetic flue gas was investigated in the
down flow reactor previously described. Two grams of the sulfide-catalyst-
material was added to the reactor and the test gas mixture passed through
the sulfide bed at a rate of 100 ml/min. The test gas contained 1000 ppm
NO , 1 percent O , 18 percent CO and the balance N . A concentration of
this magnitude would be expected in a flue gas from the burning of coal using
about 10 percent excess air. Although most flue gases probably would contain
more O then this, 1 percent was chosen because it represented the lower
limit likely to be encountered and reasonable run times could be obtained.
Sulfide-catalyst combinations that looked to be the most promising in the
electro-balance studies were investigated at 400°C. Mixtures of BaS, CaS,
SrS and FeS with NaF, NiCl . CoCl , FeCl , and Fe O were investigated
£. £, £ £3
52
-------
using 20 wt. percent promoter mixed with the sulfide. NO analysis was
accomplished using the chemiluminescent analyzer and O was analyzed using
the molecular sieve column in the chromatograph.
The results of these tests are shown in Figures 19 through 38. As can
be seen by comparing these figures, the various sulfide-promoter mixtures
behaved quite differently under the test conditions but most important in
certain cases NO removal was at a high level for significant periods of time
even in the presence of 0 . Of the mixtures tested, CaS-FeCl , SrS-NaF,
SrS-FeCl2, FeS-NiCl2, BaS-FeCl , and FeS-FeCl , appeared to be the best.
The capacity (defined as the weight of NO reduced per unit weight of
sulfide initially present from the start of the run until the concentration
exceed 600 ppm) was calculated for each combination from the data presented
in Figures 19 through 38. This gives a quantitative indication of how good
the mixtures are for NO reduction. These are listed in Table 15.
x
TABLE 15. THE CAPACITIES OF THE METAL SULFIDE-CATALYST MIXTURES
AND OF THE METAL SULFIDES
Total wt. of
Metal
Sulfide
CaS
SrS
BaS
FeS
NaF
18.6
31.7
6.3
1.9
NiCl
^
5.5
1.4
4.8
13.6
Wt. MeS
400°C
CoCl.
2
0.9
2.1
2.0
2.2
NO reduced to 600 ppm ,^3
initially
FeCl
f.
13.9
7.2
16.8
37.2
present
Fe 0
f. j
1.0
1.1
3.4
7.3
A J.U
Unpromoted
Metal 300°C
Sulfide NaF FeCl
2
0 0.1
0
2.3
0.5 2.0
As can be seen from Table 15, the capacities varied from 0 to 37.2x10
grams NO reduced/gram of sulfide initially present. The six best were
FeS-FeClX > SrS-NaF > CaS-NaF > BaS-FeCl > FeS-NiCl > CaS-FeCl . These
results generally agree with the data obtained on the electrobalance. The
FeS-FeCl and CaS-NaF mixtures were also tested at 300°C but both showed a
low capacity so lower temperatures were not investigated further.
A very limited study also investigated the effect of CO , SO , HO and
the O concentration on NO removal. Preliminary runs using unpromoted CaS
at various temperatures wi£h 18% CO and no CO in the test gas
indicated
negligible effects and so the effects of CO were not investigated further.
The effect of adding 0. 2% SO to the feed gas when a CaS-NaF mixture was
tested is shown in Figure 39. As can be seen, at least for this mixture,
SO appears not to effect the reduction reaction. However, the presence of
0.2 mole % water vapor dramatically decreases the capacity of CaS-NaF and
SrS-NaF mixtures for NO reduction at 400°C, as shown in Figures 40 and 41.
Figure 42 shows the effect of the amount of O in the gas stream on NO
reduction. As can be seen, increasing the amount of O also dramatically
decreases the capacity of the CaS-NaF mixture for NO reduction.
53
-------
p
w
06
O
O
z
100
90
80
70
60
50
30
20
10
NO
2 grams CaS/NaF
1*00°C
100 ml/minute
1000 ppm NO
1% 0
18* 2C02
Balance N_
I
10
TIME HOURS
15
FIGURE 19. THE PERCENT NO AND 0 REMOVED
JL C
BY 2 GRAMS OF CaS/NaF AT UOO°C
54
-------
100
2 grams CaS/NiCl
Uoo°c
100 ml/minute
1000 ppm NO
0
0
TIME
HOURS
FIGURE 20. THE PERCENT NO AND 0 REMOVED
jt t
BY 2 GRAMS OF CaS/NiClg AT 1»00°C
55
-------
s
>
CM
O
PS
O
O
!Z
100
90
80
70 -
60 -
50 -
30 _
20 -
10 _
2 grams CaS/CoCl
Uoo°c
100 ml/minute
1000 ppm NO
-. W -.
0123
TIME HOURS
FIGURE 21. THE PERCENT NO AND 02 REMOVED
BY 2 GRAMS OF CaS/CoClg AT 1*00°C
56
-------
o
o
100
90
80
70
60
50
_ 2 grams CaS/FeCl
30
20
10
100 ml/minute
1000 ppm NO
1% 0
Balance N
_L
J_
NO
_L
0123
TIME HOURS
FIGURE 22. THE PERCENT NO AND 0 REMOVED
BY 2 GRAMS OF CaS/FeClg AT UOO°C
57
-------
100
90
80
70
60
CM
o 50
pi
o
O
z
\
\
\
\
\
2 grams CaS/FepO,
UOO°C :
100 ml/minute
1000 ppm NO
1% (>„
\
\
\
-CO
Balance N,
30
20
10
\
\
\
\
\
\
\
\
\
\
0 L
TIME HOURS
FIGURE 23. THE PERCENT NO AND 0 REMOVED
BY 2 GRAMS OF CeS/Fe^O- AT lfOO°C
58
-------
0—0-
g
CM
O
O
z
90
80
70
60
50
40
30
20
10
2 grams SrS/NaF
_L
100 ml/minute
1000 ppm NO
°o
-co2
Balance N^
_1_
0 5 10 15
Tim HOURS
FIGURE 24. THE PERCENT NO AND 0. REMOVED
X £
BY 2 GRAMS OF SrS/NaF AT UOO°C
59
-------
o
w
o
en
o
i
100
90 -
70 -
2 grams SrS/NiCl
UOO°C
100 ml/minute
1000 ppm NO
1* 0
TIME HOURS
FIGURE 25. THE PERCENT NO AND 0 REMOVED
BY 2 GRAMS OF SrS/NiCl AT UOO°C
60
-------
I
100
90
80
70
60
o 50
o
o
SB
40
30
20
10
0
2 grams SrS/CoCl
Uoo°c
100 -ml/minute
1000 ppm NO
NO
_L
1.0
TIME HOURS
1.5
2.0
FIGURE 26. THE PERCENT NO AND 0 REMOVED
•in* w
BY 2 GRAMS OF SrS/CoClg AT UOO°C
61
-------
o
w
100
90
80
70
60
-------
p
w
O
z
100
90 -
80
70
60
50
40
30
20
10
NO
Balance
2 grams SrS/Fe 0
Uoo°c J
100 ml/mimite
1000 ppm NO
1* 00
TIME HOURS
FIGURE 28. THE PERCENT NO AND 0 REMOVED
BY 2 GRAMS OF SrS/FCgO AT UOO°C
63
-------
p
w
CM
OS
O
O
2
100
90
80
70
60
50
40
30
20
10
2 grams BaS/NaT
100 ml/minute
1000 ppm NO
1* 0
Balance N_
TIME HOURS
FIGURE 29. THE PERCENT NO AND 0 REMOVED
Jt ^
BY 2 GRAMS OF BaS/NaF AT UOO°C
64
-------
06
O
i
100
90
80
70
60
50
40
30
20
10
0
2 grams BaS/NiCl
Uoo°c
100 ml/minute
1000 ppm NO
1* 0
18* 2CO
Balance N^
I
2 3
TIME HOURS
FIGURE 30. THE PERCENT NO AND 0 REMOVED
BY 2 GRAMS OF BaS/NiClg AT
65
-------
o
w
100
90
80 -
70 -
60 -
\
\
\
\
\
\
2 grains BaS/CoCl
Uoo°c
100 ml/minute
1000 ppm NO
1* Op
18* 2co2
Balance N«
o 50
o
o
2
30 -
\
\
\
\
\
\
20 -
10 -
NO
O
TIME HOURS
FIGURE 31. THE PERCENT NO AND 0 REMOVED
BY 2 GRAMS OF BaS/CoClg AT UOO°C
66
-------
tA
O
100
90
80 -
70 -
60 -
50 -
40 -
30 -
20 -
10 -
2 grams BaS/FeCl
4oo°c
100 ml/minute
1000 ppm NO
0
18* Bo
Balance N
TIME HOURS
FIGURE 32. THE PERCENT NO AND 0 REMOVED
BY 2 GRAMS OF BaS/FeClg AT UOOPC
67
-------
o
w
Q
ai
o
o
z
100
90
80
70
60
50
40
30
20
10
I
2 grams BaS/Fe_0
100 ml/minute
1000 ppm NO
0
NO
_L
0123
TIME HOURS
FIGURE 33. THE PERCENT NO AND 0 REMOVED
BY 2 GRAMS OF BaS/FegO AT UOO°C
68
-------
CM
OS
o
100
90
80
70
60 -
50 -
40 -
30 -
20 -
10 -
0
2 grama FeS/NaF
1*00°C
100 ml/minute
1000 ppm NO
1* 0
3 I* 5
TIME HOUBS
FIGURE 34. THE PERCENT NO AND 0,, REMOVED
Ai £
BY 2 GRAMS OF FeS/NaF AT 1*00°C
69
-------
p
u
Q
-------
od
O
100
90
80
70 -
60
50
40 -
30 -
20 -
10 -
2 grams FeS/CoCT
100 ml/minute
1000 ppm NO
0
FIGURE 36. THE PERCENT NO AND 0 REMOVED
JL £
BY 2 GRAMS OF FeS/CoClg AT UOO°C
71
-------
5
OS
o
o
100
90
80
70
60
50
40
30
20
10
2 grama FeS/FeCl
Uoo°c
100 ml/minute
1000 ppm NO
1* 0.
18* 2CO
Balanced
1
3 5 10 15
TIME HOURS
FIGURE 37. THE PERCENT NO AND 0 REMOVED
>« C
BY 2 GRAMS OF FeS/FeCl AT 1»00°C
72
-------
CM
O
100
90
80
70
60
50
40
30
20
10
_L
2 grams FeS/Fe 0
100 ml /minute
1000 ppm NO
1? 0
CO?
Balance N
J_
3 U 5
TIME HOURS
FIGURE 38. THE PERCENT NO AND 02 REMOVED
BY 2 GRAMS OF FeS/7e20 AT UOO°C
73
-------
g
o
100
90
80
70
60
50
30
20
10
I
2 grams CaS/NaF
Uoo°c
100 ml/minute
1000 ppm NO
1* 0
18* 2CO
Balance N
0.2* S0
I
5 10
TIME HOURS
15
FIGURE 39. THE EFFECT OF SOg IN THE FEED GAS
STREAM ON N0x REMOVAL AT UOO°C WITH CaS/NaF
74
-------
100
90
80 U
70 U
B 6o
1
50 |-
40 |-
30 I-
20 U-
10
0
2 greuns CaS/NaP
1*00°C
100 ml/minute
1000 ppm NO
FIGURE
10
TIME HOURS
THE EFFECT OF H-O
BY CaS/NaF AT 1|00°C
ON NO REMOVAL
75
-------
Q
W
I
100
90 L-
80 |-
70 [-
60 |-
50
30
20 |-
10 1-
0
2 grains SrS/NaF
Uoo°c
100 ml/minute
1000 ppm NO
1% Op
18% 2CO.
Balance N
• 2 mole % HO
15
5 10
TIME HOURS
FIGURE 41. THE EFFECT OF HgO ON THE REMOVAL
OF NO BY 2 GRAMS OF SrS/NaF AT kOO°C
^fe
76
-------
100
90
80 -
70 -
g 60
1
50 -
i
2 grams CaS/NaF
Uoo°c
100 ml/minute
1000 ppm NO
18* CO
Balance N
30 -
20 -
10 -
0
FIGURE 42.
5 10
TIME HOURS
THE EFFECT OF 00 CONCENTRATION ON NO REMOVAL
77
-------
Finally, two runs investigated the removal of NO from the synthetic
flue gas by 5% CaS (only) impregnated on the high surface supports Harshaw
1602-T and Linde TM-0-1114. The results are shown in Figure 43. As can be
seen, performance curves of the pellets are much better than the best
sulfide-promoter mixtures in bulk form. The capacity of the pellets for NO
reduction was calculated to be 0.76 and 0.91 grams NO/grams initial sulfidex
in the pellets for the Harshaw and Linde pellets respectively, over 20 times
that of the best bulk sulfide-catalyst mixture.
Discussion
Of the 19 metal sulfides tested for NO reduction only eight (BaS, CdS,
CaS, Cu S, PbS, SrS, ZnS, and NiS) showed a weight gain. Of these, the
alkaline earth sulfides (CaS, BaS, SrS) were thought to have the most prom-
ise because of the stability of the sulfides and corresponding sulfates.
Other factors being equal, CaS would be the most desirable sulfide because
of its potential abundance and low cost. It can readily be produced from
gypsum by reduction using a number of reducing agents including coke, carbon
monoxide and hydrogen (15,16). Thus, it would be possible to regenerate the
CaS, and this would result in a process where the reducing agent was coke or
carbon monoxide:
CaS + 4NO •* CaSO, + 2N0
4 2
CaS04 + 2C -*• CaS + 2CO2
Alternatively, since gypsum is so abundant and cheap it may be econom-
ical to have a throw away process where CaSO is not regenerated.
Because of these considerations, much of the research effort concen-
trated on using CaS as the reducing agent. However, some of the data in the
research program indicated that SrS may be superior to CaS when reducing NO
in the presence of O . In addition, FeS was investigated further because of
its abundance and high rate of reaction even though SO was formed in the
reduction of NO by FeS.
The rate of reaction is significantly increased by intimately mixing
certain materials with the sulfides. K FeF was one of the most active pro-
moters but electrobalance studies showed that it reacted with NO at tempera-
tures between 200 and 500°C resulting in a weight gain. CoCi2 gained weight
when exposed to NO at 300°C but lost weight at 400°C. Fe2°3 gained weight
at 300°C. NaF, NiCl and FeCl2 did not react with either NO or O2 at 300
and 400°C. These materials significantly increased the rate of reduction
with some of the sulfides at these temperatures and thus the action appears
to be catalytic in nature. The electrobalance study showed that FeO mixed
with FeS, SrS and CaS significantly increased the rate of reaction with NO in
the absence of O but results using the synthetic flue gas containing 1% O
were poor with the sulfide-Fe O mixtures. This indicates that O_ may inter-
fere with any catalytic activity exhibited by Fe.O under these conditions.
Other trends of the sulfide-catalyst mixtures obtained during the simulated
flue gas generally agree with the data obtained in the electrobalance study.
A very active form of sulfide is obtained by impregnating it on a high
surface area support but these are so active that there is a rapid oxidation
78
-------
i
100
90 -
80
70
60
50
40
30
20
10
100-ml/minute
1000 ppm NO
1* Op
18* 2CO
BalanceTJ-
J.
2 ml. Harshav 1602-T
2 ml, Linde TM-0-
lllU
5 10
TIME HOURS
15
FIGURE 43. THE REMOVAL OF N0x BY HARSHAW PELLETS
AND LINDE MOLECULAR SIEVES IMPREGNATED WITH CaS
79
-------
of the pellets in air at room temperature. If the sulfide pellets were used
in a flue gas environment containing limited amounts of O , NO reduction
would be significant but the pellets would also be rapidly consumed with the
O . The selection of the support material would be very important since sup-
ports of different compositions behave differently. From the limited data
obtained, it appears that activated alumina containing very little silica
would be the best.
The reaction of the sulfides with O is very important in determining
the usefulness of the sulfides for the control of NO emissions from flue
V
gases. Flue gases from power plants typically would contain much more O
than NO (O /NO ratios as high as 50/1 are probable) and so when O? reacts
with the sulfide most of it would be consumed by a non-pollutant. Tests of
CaS-NaF at 400°C increasing the O concentration to 10 percent greatly
decreased the capacity of the mixture for NO reduction. However, the
electrobalance studies indicated that, for CaS-NaF mixtures at 200°C, NO
reacted at a finite rate while no reaction with 0 was detected. This could
not be verified using the CaS-NaF mixture in the tubular reactor with the
simulated flue gas, however, probably because the reaction rate was too low
to detect a change in NO concentration under the conditions of the test.
Apparently the rate of reaction is higher at the higher NO concentration
used for the electrobalance tests. x
The presence of water vapor at lower temperatures lowered the rate of
NO reduction at least for CaS. This conflicts with the earlier data which
showed HO had no effect. This indicates that the HO probably interferes
with the action of the catalyst.
SO and CO had no apparent effect on the performance of the sulfide
for NO reduction.
x
The tests using the synthetic flue gas indicate that each sulfide-
catalyst mixture behaves quite differently. The performance of the various
mixtures generally followed the trends seen in the electrobalance studies
but there were some inconsistencies with the FeS-FeCl system. The electro-
balance studies showed that FeS-FeCl had a high rate of weight gain at
300°C but that the mixture lost weight at 400°C using 2.5% NO in He. How-
ever, in the synthetic flue gas tests the FeS-FeCl had the highest capacity
for NO reduction of any of the mixtures at 400°C out the capacity was quite
low at 300°C. At the low temperature mass transfer may be limiting the
reaction rate while at 400°C the rate of reaction is limiting.
FeS forms SO when it reacts with NO and thus it would probably be un-
suitable for NO control although the rate and capacity is quite high at
400°C. X
The capacity (defined as the weight of NO reduced per unit weight of
sulfide initially present from the start of the run until the exit concen-
tration exceeds 600 ppm) was quite low for all of the bulk sulfide-catalyst
mixtures under the conditions of the tests using the synthetic flue gas
test mixture. The two highest were .0372 and 0.0317 grams NO/grams sulfide
for FeS-FeCl and SrS-NaF mixtures respectively. The best CaS mixtures were
. 0186 and . 0134 grams of NO/grams sulfide for CaS-NaF and GaS-FeCl mixture
respectively. These compare with the capacity of the 5% CaS (only) high
80
-------
surface area pellets of 0.76 and 0.91 grams/NO gram sulfide in the Harshaw
and Linde pellets, respectively. Thus, it appears that the capacity of the
bulk sulfide beds is being limited by poor gas-solid contacting. For com-
parison, assuming the reaction:
CaS + 4NO -*• CaS04 + 2N2,
the maximum capacity for CaS would be 1.66 grams NO reduced/gram of sulfide.
The pellet tests indicate that the capacity can be greatly increased by
impregnating the sulfide on high surface area supports and it may be possible
to further increase it by including a catalyst on the pellets. Reaction in a
fluidized bed reactor should also greatly increase the capacity if the bulk
material were used.
The tests to date have been limited to a synthetic flue gas containing
1 percent O . This would be typical of a flue gas from burning of coal with
10 percent excess air. A more realistic O_ concentration probably would be
about 5 percent 0_ which would result from using about 25 percent excess air,
however. The capacity of the sulfides for NO reducation would probably be
reduced in the flue gas containing the higher O concentration, but it may
be possible to obtain a reasonable capacity with the proper choice of
sulfide-catalyst-support material at lower temperatures.
The study is continuing and the results will be the topic of a future
publication.
81
-------
REFERENCES
1. Environmental Protection Agency Publication NO. AP-67. Air Quality
Criteria for Nitrogen Oxide, January 1971.
2. McCutchen, G.D. NO Emission Trends and Federal Regulation. Chemical
Engineering Progress, August 1977, pp. 58-63.
3. Hopper, T.G., and W.A. Manone. Impact of New Source Performance
Standards on 1985 National Emissions from Stational Sources. EPA 450/
3-76-017 (May 1976).
4. Karp, I.N., and O.I. Golobov. Paper SM-107181, Magnetohydrodynamics
Symposium, Warsaw, 5:2875, 1968.
5. Ando, J., H. Tohata, K. Naguta, and B.A. Laseh. NO Abatement for
Stationary Sources in Japan. EPA-600/7-77-1035, Sept. 1977.
6. Henschel, D.B. Emissions from FBC Boilers. Environmental Science &
Technology 12(5):534-38, 1978.
7. Jain, L.K. , E.L. Calvin, and R.L. Looper. State of the Art for Con-
trolling NO Emission, Part I. Utility Boilers, EPA-R-2-72-072a
(Sept. 1972?.
8. Hall, H.J., and W. Bartolk. NO Control from Stationary Sources.
Environmental Science Technology, 5(4):320-6, 1971.
9. Shelef, M., and J.T. Kuramer. Chemical Engineering Progress, Symposium
Series, 67(115):74-93, 1969.
10. Bartok, W., A.R. Crawford, H.J. Hall, E.H. Maury, and A. Skopp. Sys-
tems Study of Nitrogen Oxide Control Methods for Stationary Sources.
Clearing House for Federal Scientific and Technical Information,
PB 184 479, May 1969.
11. Ando, J., and H. Tohata. Nitrogen Oxide Abatement Technology in Japan
—1973. EPA-R2-73-284 (June 1973); NTIS PB-22 143.
12. Faucett, H.L., J.D. Maxwell, and T.A. Barnett. Technical Assessment of
NO Removal Processes for Utility Application. EPRI AF-568, EPA-600/
7-77-127, Office of Research and Development, U.S. Environmental
Protection Agency, Nov. 1977.
13. Fulton, C.H. Metallurgical Smoke. Bulletin 84, Dept. of the Interior,
Bureau of Mines, 1915, pp. 72-77.
14. Young, S.W. The Thiogen Process for Removing Sulfur Fumes. Trans.
AIChE, 8:81-89, 1915.
82
-------
15. George, D'Arcy, L. Croker, and J.B. Rosenbaum. Current Research on the
Production of Sulfur from Gypsum at the Salt Lake City Metallurgy
Research Center of the U.S. Bureau of Mines, March 1968.
16. Zadick, T.W., R. Zavalita, and F.P. McCandless. Catalytic Reduction of
Calcium Sulfate to Calcium Sulfide with Carbon Monoxide. Ind. Eng.
Chem. Process Res. Develop, ll(2):283-7, 1972.
17. Berty, J.M. Reactor for Vapor-Phase Catalytic Studies. Chemical
Engineering Progress, 70(4):78-84, May 1974.
18. Carberry, J.J. Designing Laboratory Catalytic Reactors. Industrial
and Engineering Chemistry, 56:39, Nov. 1974.
83
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
i. REPORT NO.
EPA-600/7-78-213
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Reduction of Nitric Oxide with Metal Sulf ides
6. REPORT DATE
November 1978
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
F. P.McCandless and Kent Hodgson
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING OROANIZATION NAME AND ADDRESS
Montana State University
Department of Chemical Engineering
Bozeman, Montana 59717
10. PROGRAM ELEMENT NO.
1NE624
11. CONTRACT/GRANT NO.
Grant R800682
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final; 6/72 - 5/78
14. SPONSORING AGENCY CODE
EPA/600/13
15. SUPPLEMENTARY NOTES TERL_RTP project officer is J. David Mobley, Mail Drop 61, 919/
541-2915.
16. ABSTRACT The repOrj- gjves results of research to determine the technical feasibility
of using metal sulfides for the chemical reduction of NOx to N2. Nineteen different
metal sulfides were investigated, using a test gas of pure NO. Although most sul-
fides resulted in some NO reduction, BaS, CaS, SrS, and FeS were the most pro-
mising. Several catalysts reduced the temperature at which the reduction process
proceeds by as much as 200 C. A further temperature reduction was obtained by
impregnating the sulfide and catalyst on high surface area supports of activated
alumina or molecular sieves. The most promising catalysts were NaF, NiC12, and
FeC12. All combinations of the most promising sulfides and catalysts were tested for
NO reduction, using a synthetic flue gas containing 1000 ppm NO and 1% O2. The ca-
pacities of the six best were FeS-FeC12 > SrS-NaF > CaS-NaF > BaS-FeC12 >
FeS-NiCl > CaS-FeC12, and ranged from 0. 0372 to 0.0134 g NO reduced/g initial
sulfide present. Capacities of 0.91 and 0.76 g NO/g sulfide were obtained when using
5% CaS (onlyjImpregnated on alumina and molecular sieves, respectively. It was
concluded that these sulfides caiTreduce NO in the presence of O2, but more
research is required to establish the economic feasibility.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COS AT I Field/Group
Pollution Catalysis
Nitrogen Oxides Aluminum Oxide
Reduction Absorbers
Barium Inorganic Compounds
Calcium Inorganic Compounds
Strontium Compounds
Pollution Control
Stationary Sources
Metal Sulfides
Activated Alumina
Molecular Sieves
13B
07B
07C
07D
11G
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (Thlt Report)'
Unclassified
21. NO. OF PAGES
94
20. SECURITY CLASS (TMlpage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
84
------- |